The Syntax of Zhyler
What a daunting noun phrase that is. To be honest with you (since that's easiest, according to Grice), I'm not completely convinced that syntax is even a thing, no matter how you define "thing" (or "a", or "is"...). Nevertheless, Herr Chomsky seems to think it's a thing, and since language was (officially) created by Chomsky, a discussion of syntax seems necessary.
Since syntax is such a big thing (operating under the assumption that it is, in fact, a thing, for the time being), it's not clear to me how exactly to go about discussing it. As such, I think what I'll do is start out with a simple sentence and go from there, explaining how sentences work, and so forth. We'll see how it goes.
Introduction
To begin with, I want to introduce some terminology. There are a whole bunch of noun cases in Zhyler, and you can find out more about them in the section on noun cases. For our purposes here, a noun case is a bit of phonology attached to the end of a noun. The definition of a given case does not necessarily predict its semantic role in the sentence.
Aside from terms like "noun", "verb", "adjective" and "adverb", some important terms to know will be "case marker", which is the suffix attached to a noun or adjective to indicate case, "tense marker", which is the suffix attached to a verb to indicate tense, and "modal", which is a suffix attached to a verb to indicate a whole bunch of different things.
Getting specifically to some syntactic issues, a constituent is a string of words that are treated as a single unit by a sentence. For example, in the sentence, "A good book is hard to read", "a good book" is a constituent, as is "to read", and "hard to read", and "is hard to read". "Good book is", however, is not a constituent. The notion of a constituent comes from syntaxes (there's a word you don't see everyday...) which can be represented by trees. Why trees? I bet Almighty Chomsky probably doesn't even know. It's based on an outmoded psychological theory from the 19th century that suggested that all things can be divided in two. Even though the way that trees have been used is suspect, to say the least, the notion of a constituent and syntactic trees themselves can still be useful for expository purposes, so I might use them. The important thing to remember about a constituent is that the words in a constituent belong together for some reason. In "A good book is hard to read", there's no reason to suggest that the words in "good book is" should belong together, to the exclusion of the other words in the sentence.
Another important term is the term phrase. A phrase is a string of words that have a head and (optionally) modifiers. So, take our example "A good book is hard to read". You could divide the sentence into two parts: A thing and a comment about that thing. Those two things are phrases. The first phrase is "a good book". That's called a noun phrase, because it's headed by a noun. The nominal head is "book", and the modifiers are "a" and "good". The comment, or verb phrase, is "is hard to read". Within that verb phrase is another phrase, "hard to read", which could be described as an adjective phrase. Anyway, the notion of phrase will be very important, because, as you'll soon see, Zhyler is nothing more than a series of phrases.
That'll do it for the intro. If more is needed, I'll add more later. For now, let's start off with a sentence. What sentence shall I choose...
A Simple Sentence
Here's a sentence:
Nawmos baz.
This sentence is almost as simple as you get. This sentence means "The fish laughs", or "The fish is laughing" (whatever gets you a simple present tense reading). The first word means "fish" and the second "laughs". The first word is a noun and the second is a verb. Additionally, the first word comprises an entire noun phrase, and the second word an entire verb phrase. Together, these make a sentence or clause.
Before going any further, I want to add a level of abstraction that will become useful later. As I said, Zhyler has a lot of noun cases. As such, no noun can appear without a case. In the example above, the noun nawmos, "fish" is in the default case, which is the nominative case. Schematically, we could represent this as "fish-NOM." (I put the case marker after the noun because all other case markers in Zhyler are suffixes.) Now, the level of abstraction I want to add is this: Every noun phrase (NP) is actually a part of a larger phrase called a case phrase, or CP. The case phrase is headed by the case marker, which in the case of a nominative NP, would be absent. Thus, we can have a rule that looks like this:
CP VP > S
In the rule above, S stands for "sentence", CP stands for "case phrase", and VP stands for "verb phrase". The rule above states that if a CP is followed by a VP, then the two are conjoined to form an S. This can be represented graphically with a tree. Here's an example:
The example above shows that the noun "fish" becomes a noun phrase (NP), which in turn becomes a case phrase (CP). The verb "laughs", in turn, becomes a verb phrase (VP), which in turn combines with the CP to form a sentence (S). In this way, the nominal meaning combines with the verbal meaning to produce a clausal meaning, and you get the whole sentence: "The fish is laughing." What does that signify? The answer is of less importance than the question.
Well, so far we have a basic grammar. The grammar consists of one rule which says that you can take any verb and put it after any noun and get a sentence, provided both the noun and verb are in their default state. Before we add complexity to the basic framework, let's see what we can do with this simple framework.
Expanding Upon Our Simple Sentence
So, what about nouns? They can do a bunch of stuff. For example, you can say things like, "that guy", or "a blue book", or "a self-destructing automatic pencil sharpener". In each of these, the noun ("guy", "book" and "pencil sharpener") is the head of the phrase, and everything to the left is a modifier of one kind or another. In Zhyler, nouns can be modified in a similar way. (However, there is no equivalent for "a(n)" or "the" in Zhyler. Definiteness is determined in other ways.) So, for example, if you wanted to say "the blue fish", you'd say...
Ayla nawmos
Now let's say you wanted to say "the small blue fish". Now you'd say...
Milip ayla nawmos
And now let's say you wanted to say "this small blue fish". Then you'd say...
Le milip ayla nawmos
Thus, just like in English, you can modify a noun with certain words. Also like English, there are rules of combination. So, just as you can't say, "Blue this small fish", you can't say *Ayla le milip nawmos in Zhyler. This is because le, meaning "this", is what's known as a determiner. "Determiner" is just a fancy word for the last thing that can modify a noun phrase. Examples in English are "the", "a", "an", "this", "that", "some", etc. Zhyler has a few, but not as many. One such is le, "this". So, in Zhyler, you can have a determiner, followed by any number of adjectives (which I'll label "A"), followed by a noun. Thus, we can come up with a rule that looks like this:
(D) (A+) N > NP
The rule above states that a noun phrase (NP) can consist of at least one noun, which can be preceded optionally by any number of adjectives, which can, in turn, be preceded optionally by at most one determiner. We can call this our NP rule.
Now, let me point out some differences between English and Zhyler. First of all, in English a noun must always be accompanied by a determiner, unless that noun is in the plural or is a mass noun. In Zhyler, a noun by itself is enough, because nouns are self-determining.
Another difference between Zhyler and English is that adjective order is fairly free in Zhyler, whereas it isn't in English. In English, you can say "the small blue book", but it's really bizarre to say "the blue small book", unless "small book" was understood to be some sort of compound. In fact, messing with adjective order pretty much forces a compound reading in English. Not so in Zhyler. In Zhyler, the most important adjective is put first, and the rest strung afterwards. So, if you wanted to say "this small blue fish" (to stick with our example), you'd say, Le ayla milip nawmos, whereas if you wanted to say, "this small blue fish", you'd say, Le milip ayla nawmos. This sort of arrangement is paralleled directly at the clausal level, but we'll get to that later.
So, for right now, we still have our basic sentential rule, and a complex NP rule. Here's a tree of what our expanded simple sentence looks like:
For right now, we'll work with a four-branch tree. Later we may want to emend the analysis, but it's fine for now. One important thing we can draw from this right now, though, is how the tree structure reflects what we've found about the adjectives. Namely, it's not the case that the noun forms a constituent with the first adjective, and then this phrase forms a constituent with the next adjective. This is because the ayla nawmos in the sentence Le milip ayla nawmos baz does not mean the same thing as ayla nawmos in the sentence Le ayla nawmos baz. This should be reflected in the tree structure, which is why you have separate branches for each adjective.
Adding Adverbs
Before adding extra case arguments, let's see what we can do with that verb phrase. So, in English you can say, "This little blue fish is laughing", but you can also say "The little blue fish is laughing loudly", or "The little blue fish is laughing now". You can do the same in Zhyler. Both "loudly" and "now" are adverbs in Zhyler, though they're adverbs of two types: Adverbs of manner and adverbs of time. They're done a little different, but they can be added now without causing too many problems.
Starting with adverbs of manner, the word for "loud" in Zhyler is jet. To make it into an adverb, you add the adverbial adjectival case suffix, which, in this case, would be -e. Thus, jete is "loudly".
The traditional adverbial position in Zhyler is preverbally, and adverbs of manner adhere strictly to that tradition. So, to say "This little blue fish is laughing loudly", you'd say:
Le milip ayla nawmos jete baz.
Placing the adjective anywhere else will result in robust ungrammaticality.
Next comes adverbs of time. Temporal adverbs enjoy a little more freedom. Their usual spot is right in front of the verb, as with adverbs of manner. So, if you wanted to say, "This little blue fish is laughing right now", then, using the Zhyler word for "now", ye, you could say:
Le milip ayla nawmos ye baz.
Additionally, though, a temporal adverb can be moved to the front of the clause. This is done to focus the timeframe being used in the clause. You can compare this to English temporal adverbs being used sentence-finally vs. sentence-intiially. If you say, "Yesterday, I went to the store", the timeframe is focused in a way that it isn't in "I went to the store yesterday". Here's an example of a fronted temporal adverb in Zhyler:
Ye le milip ayla nawmos baz.
When a temporal adverb is fronted, it belongs to a different intonational unit than the rest of the sentence. So, in the sentence above, ye belongs to one intonational unit, and the rest of the sentence to a second (and, of course, there are intonational subunits within the sentence itself). And, since ye is our adverb of choice for the moment, you might as well know that ye often shows up reduplicated as yeye when it's used sentence-initially. Here's an example:
Yeye, le milip ayla nawmos baz.
So let's see how this affects our grammar. We now know that a verb phrase can consist of an adverb plus a verb, so we can add this rule:
AV V > VP
I'm going to use AV as an abbreviation for "adverb", for the time being. So that's fine. But what about for the fronted adverb? It wouldn't be a good idea to include that as part of the noun phrase, since it affects the meaning of the whole clause. It also wouldn't be a good idea to say that, for example, AV CP VP > S, because that would make a fronted adverb look like a crucial part of the sentence, which it isn't. For that reason, I'm going to invent a new symbol: Z. A Z will stand for a Zhyler sentence. In order to define Z, we'll need another rule:
(AV) S > Z
Though it'll be jumping ahead, I want to briefly refute a movement analysis. One might argue that adverbs are always generated in front of the verb, and that when fronted, they move to the front of the sentence. I would argue that fronted adverbs are generated right where they sit in the sentence, and that they're less time adverbs, than time reference markers. So, for example, take the following sentence:
Šeyajamey, milip ayla nawmos surwÿruzðaz bazlar.
The sentence above, translated word for word, is "Yesterday, small blue fish at noon laughed". In English, the equivalent would be, "Yesterday the small blue fish laughed at noon". The point of showing this is to show that the first adverb sets up the overall timeframe, and the second specifies a point within that time frame. Thus, the two different adverbs are performing different functions, and it's not the case that the fronted adverbs undergoes some sort of movement operation.
Back to our simple sentences, here are some trees of the two sentences above: One with an adverb in preverbal position, and one with a fronted adverb:
These trees will work for now. The thing the trees will help show you is what is and is not a constituent. As you can see with the tree on the right, yeye doesn't form a constituent with anything, except for the sentence as a whole.
Getting Argumentative
Keeping all the previous tidbits in mind, let's make things a little more complex. So far, all we've dealt with are intransitive verbs, that is verbs that only have one argument. In each case, the argument has been the subject of the sentence. By adding arguments, though, we can what are called transitive verbs, or verbs with both a subject and object. So, what's a good one... How about "to see". That's a good one. So, when you see something, you, in a way, see something. That something is what I'm going to call the direct object of the sentence. [Note: If this were a different language, like Epiq, or Gweydr, you might want to call it something other than an object. For Zhyler, though, object is appropriate.] The direct object is an argument that is at the opposite end of an action from the subject. So, on one end of seeing is the seer, which is the subject, and on the other end is the seeee, which is the object. [Ha! Look at that: Four "e"'s in a row! And I think it's correct, too...] In Zhyler, the case most commonly assigned to direct objects is the accusative case. Here's the sentence "The fish sees the rabbit" in Zhyler:
Nawmos koyvener mat.
As you can see above, first comes the subject, then the direct object, and then the verb. This gives us the dominant word order of Zhyler: SOV. Unlike Latin, where SOV was a preferred word order, the word order in Zhyler is rather staunchly SOV, much like in Japanese or Turkish.
In order to add to our grammar, we need to find out what the object forms a constituent with. There a number of ways to do it, but the way I'll do it now is with a question. So, if I ask, "What did the fish do?", I can say "The fish ate an apple", but I can also just say "Ate/eat an apple". Now imagine I asked, "What happened to my apple?" The full answer would be "The fish ate an apple", but you could never say "*The fish ate". This is a clue that the direct object forms a constituent with the verb, but that the subject doesn't. The same is true of Zhyler, though I can't use this example (an omitted subject or object is interpreted as a discourse relevant third person subject/object). Thus, we can add the following rule:
CP V > VP
However, we already have a VP rule. Recall:
AV V > VP
Therefore we need to try a sentence with a transitive verb and an adverb. So, here goes:
Nawmos koyvener ye mat.
That is "The fish sees the rabbit right now", and, as you can see, the adverb still directly precedes the verb. Thus, we can emend our VP rule to look like this:
(CP) (AV) V > VP
The CP and the AV are in parentheses above because they're optional. The only requirement for a VP (thus far) is a V, as we've seen.
So, now we've gotten a little more complicated. Here's a tree to help explicate. The sentence is Le nawmos za koyvener ye mat, "This fish sees that rabbit right now":
The tree above shouldn't be hard to understand, since all that was added was an extra CP. It's important to note that the CP is dominated (i.e., it has a line going up to something else) by the VP node (a node is one of the parts in the tree that says VP or CP or NP, etc.). This is what makes that CP a direct object, as opposed to a subject. This point will be expanded upon later (by me [or an associate of mine]).
There Is Trouble in the Forest...
Thus far we have a nice, neat looking syntax. It doesn't produce very many sentences, but those that it does produce are nice and neat. Now here are a few questions which are going to cause us to totally deconstruct and rebuild our syntax, and which will really give George W. Chomsky a headache:
- What happens if you include either a first or second person argument in a Zhyler sentence?
- What happens if you use a tense other than the present tense?
- What about modal and aspectual information?
- If Zhyler has fifty-seven noun cases, why have we only seen two so far?
I'll tackle each of these questions one at a time. First on the block is question 1, regarding first and second person arguments.
Adding First and Second Person Arguments
Until now (hee, hee... A Kamakawi joke), we've only seen sentences with third person arguments (e.g., the little blue fish, the rabbit, etc.). But you know what? You and I can do things in Zhyler, too, and it's about time we asserted our presence. So, forget "The fish sees the rabbit". How about "I see the rabbit":
Koyvener matum.
The only thing changed is the suffix on the verb. To mark a first person subject, you add an -Um suffix to the verb (which, in this case, shows up as -um). So that's how you mark a subject. So how about an object? Here's "The fish sees you":
Nawmos matle.
By adding the -lR suffix (which shows up above as -le), we can express a second person object. So what about adding both: "I see you":
Matlem.
That above is a whole sentence using a second person direct object suffix and a first person subject suffix. [Note: If you want to see how the vowel harmony for these things work, go to the page on verbs.] So, what does this do to our rules and tree structure? Are both subjects and objects superfluous? Almost. Consider the following sentences:
Mat.
Indeed: That's a whole sentence. It means "She sees him", or "He sees it", or "It sees her", etc. No person suffix on a verb and no nominal arguments in a sentence mean that the verb has an unexpressed, unspecified third person subject and object (for transitive verbs) and an unexpressed unspecified third person subject (for intransitive verbs). So, for example, if you had the sentence Eren, it would mean "S/he/it's sleeping", and never "*S/he/it's sleeping her/him/it". This is because the verb ernel, "sleep" can't take a direct object.
So, what to do? Well, notice that there's something interesting about the ordering of the arguments. I said above that Zhyler's word order is rather stringently SOV. But, when you have, let's say, two argument suffixes on the verb, the word order becomes VOS. Additionally, you could have SVO or OVS (one argument expressed as a suffix; the other overtly). Given the six possible word orders, the only two left out are VSO and OSV. What do these two have in common? This: The direct object does not occur next to the verb. And there you have a generalization: In Zhyler, the direct object (or object phrases) must occur next to the verb (that is, a subject may not intervene between the object and verb). That's an interesting generalization. So, if we called the argument suffixes CP's, just for the sake of argument (excuse the
(CP), [(AV) V] > VP
(AV) [(CP), VP] > S
Let me go over what these rules say. First of all, let me explain some symbols. As I explained before, ( ) is used to indicate that an argument is optional. The brackets [ ] are there to indicate that the ordering within those brackets is fixed. This is important because the comma , indicates that the order of the things on either side of the comma can be flipped. Thus, with a VP that only has a V and a CP, you can get CP V (that would be an overtly expressed object), or V CP (that would be an argument expressed with a suffix). The same goes for the second rule. Notice that the AV still appears only sentence-initially or
All right, that (for the time being) solves the mystery of first and second person arguments. Here are some trees to help illustrate just what changes have been made. I'll do four. The first will be a sentence with the subject expressed as a suffix: Koyvener matum, "I see the rabbit". The second will be a sentence with an object expressed as a suffix: Nawmos matle, "The fish sees you". The third will be a sentence with both the subject and object expressed as suffixes: Matlem. And finally, I'll go ahead and do a tree of Mat, just so you can see what it looks like. Here they are:
My, but don't those trees look cluttered! Cluttered trees with needless levels (e.g., Z, CP) are a clue that there's something wrong with your syntax. (The Minimalist folks don't seem to have picked up on this fact.) Such is the case with the syntax we have right now. Nevertheless, it's a good idea to start out this way, 'cause things are going to change drastically in a minute. The important thing to glean from these trees is how the direct object and verb combine. So, if you look at the CP level, there are two CP nodes, and the one that combines with the verb is always the direct object. The one that combines with the VP, on the other hand, is always the subject. Keep that fact in mind as we move on.
Adding Tense
Thus far, every sentence we've seen has been in the present (or unmarked/default) tense, but there are other tenses in Zhyler. There's a specifically marked past tense (-lAr) and future tense (-zA), as well as an irrealis marker (-wW), which tends to act an awful lot like a tense. There's also an infinitive marker which nominalizes the verb (Class XIV, for those interested). These tenses are suffixes that are added to the verb, not verbal inflections, so to speak, so they get their own separate marking (in a tree, I mean), just like nominal suffixes on the verb. So here's a very simple sentence with a tense marker:
Nawmos koyvener matlar.
That's "The fish saw the rabbit". Now, once can imagine that the -lar suffix could get a category, T, and that then we could emend our VP rule to be something like:
(CP), [(AV) V (T)] > VP
And that'd be it. Right? Well, not so fast. This emended VP rule would work fine for the sentence above, and even a sentence with a subject expressed as a suffix, such as "I saw the rabbit":
Koyvener matlarum.
But problems arise the moment you try to include an object suffix. Take the following, for example:
Nawmos matlener.
Two things to note about this. First, the suffix -ler became -ner, because of a sound change that happens in Zhyler (click here to see that sound change explained). Second, there's a nominal argument on either side of the tense! This forces us to posit a really bizarre-looking VP rule:
[(CP), [(AV) V]] (T+) > VP
This rule fixes it so that the T always comes after the verb, and that an object suffix will always come between the verb and the tense suffix. This works even if you add a bunch of tense suffixes, as with this sentence-like NP:
matlenerwilüm
That means something like, "For me to have possibly seen you", or something. Anyway, -ner and -wi and -l would each be tense suffixes, and the rule we wrote above could handle it. So, bizarre though it may be, the rule works, for now. So, here's a sentence treed: Nawmos koyvener matlarwÿ, "The fish might have seen the rabbit":
Notice that there are two T's in that tree and that now our VP has a bunch of branches beneath it. We should be able to trim some of these soon.
Adding More Verbal Suffixes
Zhyler has a bunch of other verbal suffixes which come into play (you can find more info on them by going to the page on verbs). So, for example, there's the progressive suffix -mYs. This suffix emphasizes the fact that an action is ongoing. Here's an example:
Šeyajamey miyÿlimüslerüm.
That means, "I was kissing you yesterday". Currently, our VP rule can't handle this, but a very simple modification can take care of it. I'll call any verbal suffix that's not a tense suffix an AS suffix, which stands for "aspect" (I'm going to use this term very loosely). Here's how our rule could be reformulated:
[(CP), [(AV) V]] (AS+) (T+) > VP
Now comes the devastating part for proponents of the current syntax (of which there are none, in reality, so I guess it's not that devastating). Here it goes: There are verbal suffixes that can occur before the direct object suffix.
At first, this might not seem like such a problem. After all, where does one draw the line between inflection and derivation? One could easily say that all suffixes that come before the direct object suffix are derivation, and all that come after are inflection. That would give us a rule like this (using PD to mean "pre-direct object verbal suffix"):
[(CP), [(AV) V (PD+)]] (AS+) (T+) > VP
This rule could help explain a sentence like that below:
Miyÿkÿnlimüslerwim.
That sentence means, "I could have been kissing you". (To help break it up: miyÿ = "kiss"; -kÿn = "can"; -li = "you"; -müs = "be (progressive)"; -ler = "(past)"; -wi = "(irrealis)"; and -m = "I".) So you might get away with saying that the pre-direct object suffixes are of one type, and the post-direct object suffixes are of another. Still and all, though, it becomes increasingly difficult to say what's what when you consider a sentence like this:
Miyÿzirliyeswiriz!
That very long confusing word means, "Please don't get kissed by each other". (Again, broken up: miyÿ = "kiss"; -zir = "(distributive)"; -liye = "you all (indicates command)"; -s = "get (passive)"; -wi =
This tree looks pretty nasty, I'll admit (get it? "Pretty nasty" = PD N AS T T?! Hee, hee...! Ahh, it's like I did that on purpose...). Worry not, though: Salvation is on the way!
Adding More Case Phrases
The final question has to do with other noun cases. Zhyler has fifty-seven noun cases, but we've only seen two so far: the nominative and the accusative. There are others. There are even adjectival cases. So how, exactly, do they fit into this framework we've created? Well, that question will turn out to be the straw that breaks the camel's back. But worry not! We're going to nurse the camel back to health and ride him to Mecca. All will be resolved. First, though, we must go through the painful part.
Above we said that case is determined by what combines with what. So the CP that combines with the V gets the accusative case, and the CP that combines with the VP gets the nominative case. That's fine. So what about ditransitive verbs? Well, as it turns out, certain types of ditransitive verbs can work in our system. Let's take the verb elen, "to give". Even though Zhyler's cases will determine what role each CP plays in the sentence as a whole, the word order is fixed. So, if I wanted to say, "The fish gave the trophy to the rabbit", I'd say:
Nawmos koyvenles eleščer eller.
In the sentence above, nawmos, "fish", is in the nominative, elešče, "trophy", is in the accusative, and koyven, "rabbit", is in the dative. Crucially, the following is ungrammatical in Zhyler:
*Nawmos eleščer koyvenles eller.
The meaning is recoverable, but it sounds something like, "I give present you" for "I give you the present". You can figure out what it means, but it sounds like the person who uttered it doesn't speak English that well. Same thing with Zhyler. That's good for our current theory, though, because we can just say that the sentence gets bracketed like this:
[ Nawmos [ [ koyvenles [ [ eleščer eller ] ] ] ] ]
You get that? In other words, the direct object eleščer combines with the verb eller to make a VP. We could then say that the CP koyvenles combines with the VP to form another VP. We could then say that the CP that combines with the V gets the accusative case, and that the CP that combines with the first VP gets the dative case. We might want to fix things up so that we give the first VP a different name, so that it's different from the VP that combines with the subject CP to make an S, but I don't even want to do that, because I want to bring up two damning pieces of evidence that will cause us to totally restructure our current syntax. In other words, if your leg's been severed from your body, why bother putting a bandaid on the cut on your former leg's shin?
All right, here comes the evidence:
- There's a case called the benefactive case which expresses things like the underlined phrase in the sentence, "I did your homework for you". Its marker is -š. Now here are two sentences in Zhyler which both are perfectly grammatical:
- Nawmos liš koyvener miyÿlar, "The fish kissed the rabbit for you".
- Nawmos koyvener liš miyÿlar, "The fish kissed the rabbit for you".
- There are two adjectival cases which do nothing more than agree with the noun they modify. No marker at all is the nominative adjectival case marker and marks adjectives which modify a noun in the nominative nominal case. The marker -B is used for the nonnominative adjectival case marker. This case marks any adjective that modifies a noun in any other case than the nominative nominal case. So, for example: Milip nawmos milipÿ koyvener matlar, "The small fish saw the small rabbit". Now, if an NP becomes a CP after an adjective combines with a noun, how on Earth does the adjective figure out whether the noun it's modifying is in the nominative nominal case or not?
So, here we are. Our syntax has absolutely no way of explaining (1), and it's answer to (2) is an emphatic, "Beats me". Something must be done. And, rather than trying to solve the problem with crazy movement rules and selectional restrictions and base-generated arguments and case checking, and all that utter nonsense, I'm going to do something which will really get Chomsky's goat: Delinearize the tree.
Just about every framework that uses a tree-like structure (morphology and phonology and semantics, in addition to syntax has some framework that does) has one rule in common: Lines may not cross. In fact, they can't even touch in a right-side up V
Hele au i ka hale.
Since objects are supposed to form a constituent with the verb, the way most Chomskyan syntacticians would explain this is by saying that the object phrase i ka hale ("to the house") originally combines with the verb hele at some deep level that's never seen, and then in order to get the correct word order, the whole phrase is moved to the right before the sentence is spoken. (Well, they actually would probably say that the verb moves leftward, because rightward movement is kapu, for some reason.) Such an analysis ignores the fact that the most basic word order of Hawai'ian is VSO. And the only reason such an analysis was proposed was because the verb has to combine with the object before the subject, and to do so without movement would force the lines to cross.
Needless to say, I think this is nonsense. And, since this is my language and my syntax, I'm going to do away with the notion, because I don't need it. What I will propose instead is a three-dimensional tree. We only have two dimensions online (so far), so in order to achieve a 3D affect I'm going to use the space above a sentence (as we've seen so far), below the sentence, and to the right of the sentence. Thus, trees will end up looking like a diamond without a left
Before showing you any trees, though, I need to explain what motivated this radical departure from "normal" syntax.
There are two different levels of meaning in Zhyler sentences. First, there are the nouns, adjectives and verbs which cover the context (or content) of the sentence. Then, however, there's the function of the sentence. The function of the sentence deals with cases, tenses, modals and adverbs. This effectively splits both nouns, adjectives and verbs in two, and also makes a distinctive distinction between adjectives and adverbs (something that many syntaxes don't usually do). This was part of the motivation.
The other part of the motivation has to do with multiple constituencies. The rule in X syntax is that a single word (or node) can't be a part of two different constituents. In Zhyler, some words must be a part of two different constituents, because there are two levels of constituency in Zhyler: Contextual and functional. So, let's take these two example sentences in Zhyler again:
- Nawmos liš koyvener miyÿlar, "The fish kissed the rabbit for you".
- Nawmos koyvener liš miyÿlar, "The fish kissed the rabbit for you".
These two sentences mean the same thing, but are, nevertheless, slightly different. The first sentence would be considered the default order for the sentence. Thus, if you just wanted to say the sentence "The fish kissed the rabbit for you", you'd say sentence (a). In sentence (b), the direct object is focused. That is, the proper rendering into English of sentence (b) would be, "The fish kissed the rabbit for you". By putting the direct object in second position, the speaker makes that argument more prominent than the other object phrases. Sentence (a) is a neutral sentence, though, because if you wanted to focus "for you", you'd have to use a different focus position (and a different case). This is a special kind of focusing that can only occur with the leftmost core object argument of a verb. Nevertheless, in sentence (b), you have a tribranching tree: The first node under the VP is koyvener; the second is liš; and the third is the V miyÿlar. Thus, there's no way to tell which argument is the core argument and which is the non-core (or oblique) argument.
This is where the functional level comes into play. Below the sentence, the words are divided into C's (cases) and T's (tenses). Amongst the cases, there are CC's (core cases) and OC's (oblique cases). Each verb specifies how many CC's it can have, and what cases each CC gets. How a verb assigns case to its CC's is a different matter which will be taken up later. Anyway, the objective CC's group together to form a CCM (a core case module). Thus, a verbs object CC's will always form a constituent separate from everything else, and independent of the structure of the sentence. The CCM then can combine with OC's to form a CCP (core case phrase), which then combines with a T (as well as any aspectual markers that exist) to form a TP (tense phrase), which then combines with the subject CCP (it has no CCM level) to make a Z (the inverse of S, in an orthographical way). Then the Z combines with the S to form an M: a meld. In this way, contextual information melds with functional information to produce the entire clause. And that, in a nutshell, is how sentences in Zhyler work.
In words, this probably sounds confusing. I'll explain with trees. First let's revise our grammar, though.
The grammar will now be divided into contextual rules and functional rules. There will also be melding rules which deal with the interface of the contextual and functional halves of the grammar. So far, there's only one, but there will be more. I'll start out by listing the contextual rules.
So, first, our NP rule hasn't changed:
(D) (A+) N > NP
Our VP rule, though, has changed significantly:
(NP+), V > VP
The rule has been much simplified from its last manifestation, and now NP replaces CP. Additionally, I wanted to specify that there can be any number of NP's. This number will be constrained by some general rules (e.g., you can't have more than one subject unless they're conjoined, etc.), but the idea behind it generally holds true. Also, as you can see, AV no longer participates in the contextual realm: It's been banished to the functional realm. [Note: Suffixal pronouns are considered NP's, and are covered by this rule.]
Finally, our S rule has been greatly simplified:
(NP) VP > S
Those familiar with the history of syntax will look upon this rule with nostalgia. (Indeed, remember the good old days, when there were no such things as empty operators, theta roles, and A movement...? I know Chomsky wants to.) Note that the subject NP is optional. Remember that an absence of NP's will be interpreted as a non-specified third person argument (this goes for both subject and direct object).
And that does it for the contextual rules. Now for the functional rules... These are going to take some explaining. Thus, I'm going to devote an entire section to them, and go slowly, step-by-step.
Functional Rules
The functional rules are rather complicated, so I'm going to start out slow. To begin with, we're going to abandom the contextual realm, for now. After the functional realm is complete, we can marry it to the contextual realm, and see how the whole system works in concert.
Let's go back to some of our simple sentences. Here's one:
Nawmos baz.
Below the equator, this sentence is comprised of case and tense. The tense, of course, is present ("The fish is laughing" or "The fish laughs"), and the case, nominative. A verb gains its tense information either by default, or with an overriding, overt tense marker. A verb with no overt tense marker is always interpreted as being in the default tense, which, in the sentence above, is the present tense. As for the case of the subject NP, well, we have no way of saying right now how it gets its case. That will come in time. Suffice it to say that the following rule holds:
(AC+) CC > CCP
That rule states that a CC becomes a CCP, and can have as its sister any number of optional adjectival cases (AC). That takes care of sentences such as that below:
Milip nawmos baz.
This insures that milip will be in the adjectival case that corresponds to the case that the NP to its right is in.
To get back to tense (T), a T forms a TP. All it needs to go to the TP level is a T, but it can combine with other things. So, for example:
Nawmos jete baz.
That's, "The fish laughs loudly". Such adjectives directly precede the verb or T, so we can now add a TP rule that looks like this:
(AV+) T > TP
Notice how this rule looks a lot like the CCP rule. This is an example of how the nominal and verbal parts of the sentence mirror each other. Anyway, now let's add an object:
Nawmos koyvener jete baz.
Now we can simply emend our rule to include an optional object:
(CCP) (AV+) T > TP
And now we're getting pretty close. One final rule we need is a method of combining the intitial CCP and the TP. This is where our old friend Z comes in:
(CCP) TP > Z
And now we have a nice mirrored structure. Here's a tree (so far) for the sentence Le ayla nawmos za milip koyvener jete miyÿlar, "This blue fish quickly kissed that small rabbit":
I realize that this tree may look strange for a number of reasons. One reason could simply be because of the 2D mode of representation we're stuck with on the internet. To help those who can't get passed the 2Dness of this tree, here's the same tree as is shown above flipped 90 on along the Y axis. Thus, up is still up; down is still down. What was right, though, is now facing you (coming out of the screen!), and what was left is now shooting out the back of your screen. Ideally, a fully 3D model would be best for showing this kind of "tree" (which, by now, no longer looks like a tree in any way), but since we've only got 2D, I hope this inverted tree will give you a clearer picture of what's going on. Here it is:
So, compare the tree above to the one just above that one. I've pointed out where the sentence level is and the three main nodes. The lines in between represent the different node levels of the tree (i.e., where the N's and V's are, where the AC and CC's are, etc.), and the long lines represent the connecting lines. So that's what our trees look like now. It's important to understand how the structure works, because now we're going to get into the details of the functional level, and also add some depth to the contextual level.
Can Our Grammar Handle It?
It's time to play the game that every syntactician loves: Can Our Grammar Handle It? In this game, we come up with some sentences and see what our grammar does with them. Ideally, a tree would follow each sentence, but I don't want to add too many pictures to this page, for the load time will suffer.
So, first up: The simplest sentence we've examined so far:
(1) Nawmos baz.
Can our grammar handle this sentence? Yes. On the contextual level, this is simply an NP combining with a VP to form an S. On the functional level, it's a CCP combining with a TP to form a Z. Then the Z combines with the S to form an M. Simple as that. Next sentence:
(2) Koyvener matlar.
This is a sentence without an overt subject. Can our grammar handle it? Yes. In this case, the NP combines with the V to form a VP, which then goes to an S. On the functional level, the CC becomes a CCP which combines with the T to form a TP, and this TP goes to Z, and the Z combines with the S to make an M. Done. Next:
(3) Nawmos koyvener matlar.
Can our grammar handle this sentence? Yes. The steps described for sentences (1) and (2) are simply combined. Next:
(4) Bazum.
A new trick. This sentence means, "I'm laughing". Can our grammar handle it? Yes. The suffixal pronoun -um simply counts as an NP. The V baz goes to a VP and then combines with the NP to the right to form an S. Below, something needs to be explained. Suffixal pronouns do not get case. Therefore, though a suffixal pronoun is an NP above, it corresponds to nothing below. Thus, the T goes to a TP which goes to a Z, and then the S and Z combine to form an M. Next:
(5) Matlener.
Can our grammar handle this? Yes. It's not that different from sentence (4). In this case, the object suffixal pronoun combines with the V to form a VP, which then goes directly to S. Below the same thing happens as with sentence (4). Next:
(6) Matlenerüm.
Can our grammar handle both in one word? You bet. The V combines with the NP immediately to its right to form a VP, and the VP combines with the remaining NP to make an S. Below, the T goes to TP which goes to Z, which combinse with the S to make an M. Not too tough. Next:
(7) Nawmos matlener.
Can our grammar handle this mixture? Sure. The V simply combines with the suffixal pronoun to make a VP which combines with subject NP to produce an S. Below, the T goes straight to TP, and the CCP to the left combines with the TP to form a Z. But a question: The verb matal, "to see", has at least two core cases it can dole out: the nominative and the accusative. How does it know to give out the nominative to the first NP rather than the accusative? We shall have to wait and see. First, another sentence:
(8) Koyvener matlarum.
Can our grammar handle it? Yes, but what about the case assignment? How does it know to give the NP to the right the accusative? Again, we shall have to wait. (Note: In the above sentence, the V combines with the NP to the left and then combines with the subject suffixal pronoun to the right.) Next:
(9) Nawmos koyvenles eleščer eller.
Can our grammar handle this? No, not yet. In order to beef our grammar up so that it can handle a sentence like this, we'll need to add some complexity to the functional level. And, while we do so, we can answer some questions along the way.
Working with the Functional Level
Remember above when I mentioned CCM's? They will now come into play. As I mentioned above, the core cases have a strict ordering, such that the following are facts of Zhyler:
- Nawmos koyvenles eleščer eller.
- *Nawmos eleščer koyvenles eller.
Thus, the indirect object must come before the direct object. This ordering (between indirect object and direct object) is fixed across all the verbs of Zhyler. The ordering of other core cases that verbs make use of is dependent on the verb (though general tendencies can be illustrated). When an oblique case (OC) is introduced, though, some variant ordering can occur, as with the following (also, note sentence e):
- Nawmos liš koyvenles eleščer eller.
- Nawmos koyvenles eleščer liš eller.
- *Nawmos koyvenles liš eleščer eller.
Thus, we need to emend our grammar such that the ordering of core cases remains fixed, but the ordering of core cases with respect to oblique cases is fluid (but not too fluid, as can be seen with sentence e).
The answer to this problem is too introduce another level: The Core Case Module. The Core Case Module (CCM) is a module wherein all core cases are held. This module will only be relevant to the object core cases, and not to the subject core case. (Well, it is there, in the case of the subject CC, but it doesn't do anything. Yet.) Inside the CCM, CC's are ordered left to right in order of least importance to greatest importance, according to the verb. That's how the verb assigns cases. It goes to its first CCM (structurally) and assigns cases from left to right. So, in the above example, the ordering of core cases is: (1) dative; (2) accusative; and (3) nominative. So first the verb assigns the dative case to the first CC it comes across in the CCM, which is koyven, "rabbit". It then goes to the next NP, elešče, and assigns it the accusative case. Finally, it looks for the last CCM, which is the subject CCM. It's last structurally because it's the least embedded CCM. It then looks for the only NP in the subject CCM and assigns it the nominative case. But anyway, our CCM rule looks something like this:
CC+ > CCM
But, oh, there's one little troubling bit left over. After all, there are still adjectival cases to be accounted for (AC's). We want to make sure that AC's aren't just lumped in with the other CC's, because each CC and its associated AC forms a constituent of its own. Therefore, we need to invent an intermediary level which we'll call the "CC bar level", that looks like this (and, yes, we'll need this level elsewhere):
(AC+) CC > CC
So now our CCM rule looks like this:
CC+ > CCM
So that's the CCM level. But beyond the CCM level is the level we already know about: the CCP level. The CCP level is comprised of the CCM plus whatever other OCP's (oblique case phrases) there are to the right and left. We can write a simple rule for it that looks like this:
(OCP+), CCM > CCP
It occurs to me that we haven't yet written an OCP rule. It's fairly straightforward. This is what it looks like:
(AC+) OC > OCP
An OCP doesn't need a bar level because OC's don't cluster together. No, each OCP stands alone. Desperado. Anyway, cases aren't assigned to OC's. This is because OC's are self-defined. That is, when someone wants to express something using an OC, s/he has to choose the case in order to express the concept. With core cases, the verb is the one that chooses the cases; the speaker just chooses the "semantic" roles (e.g., subject, direct object, indirect object, etc.). Also, since an OC is always going to be a case other than the nominative, its associated AC will always be in the nonnominative. So that, at least, is simple.
All right, we've seen a lot of theory for awhile, so here's a tree to help explicate. This is a tree of the sentence Nawmos koyvenles eleščer liš eller, "The fish gave a trophy to the rabbit for you". For right now, I'm only going to show the functional level of the tree. Thus, it'll go up to Z, not M. This is because we're going to need to discuss the contextual level for sentences like this one, so I don't want to put up an incorrect version of the contextual level with a correct version of the functional level. So, here's the tree:
So that's how that tree works. This should give us a pretty good understanding of how the functional level works. Now let me switch focus to the contextual level so we can see what goes on when there are mutliple object case phrases (that is, both oblique and core), and when word order gets switched up.