AN INTRODUCTION TO SYNTAX
Within theoretical linguistics, syntax is the study of the architecture of phrases, clauses, and sentences. The modern roots of the study of syntax can be traced to the pioneering work of Noam Chomsky, who in 1957 wrote Syntactic Structures. Chomsky changed the face of linguistics by casting its domain inwardly. That is, the concern shifted from describing external language phenomena to characterizing the mental machinery that supposed to explain the native speakers' knowledge of language.
1. Dillon and Joelle went to the beach
2. [*]Dillon and Joelle put to the beach
Our intuitions about English (driven by our unconscious linguistic knowledge) allow us to effortlessly judge (1) as acceptable (i.e., grammatical, well formed) and (2) as unacceptable (i.e., ungrammatical, ill formed-signified by an *). If you were asked why this is so, you might hypothesize that "sentence (1) is well formed and sentence (2) is not because they have different verbs." But such a statement doesn't really tell us anything substantial; it's just a description of the difference between the two sentences. For example, to explain the grammaticality of (1) and the ungrammaticality of (2), we would have to know something about how we acquire the lexical category of verbs, how phrasal categories are constructed out of lexical categories, how each verb picks its linguistic environment (some verbs go with some phrases and some go with other phrases, to which both  and  attest), and how this information comes together to yield knowledge of the grammatical sentences of a language. Later, the theoretical machinery that some syntacticians claim is necessary to explain the facts about (1) and (2) will be described.
Our syntactic knowledge also allows us to go beyond making judgments about whole sentences. If (1) were to be divided into two sections, where would the division occur? It is very likely that the sentence would be divided between its subject (Dillon and Joelle) and its predicate (went to the beach). Thus, a first pass at the syntactic structure of the sentence might be:
3. [[Dillon and Joelle] [went to the beach]]
Note that the outer set of brackets marks out the entire sentence and the inner sets of brackets divide the sentence up into further parts. But why wouldn't (1) be divided into the following?:
4. [*][[Dillon and] [Joelle went to the beach]] or
5. [*][[Dillon and Joelle went] [to the beach]]
Sentence (1) would likely be divided as (3) and not as (4) or (5) because you have the mental capacity to effortlessly know that (3) yields well-formed parts (or constituents) and (4) and (5) do not. Now take what we have called the predicate and divide it further, into two parts:
6. [[went] [to the Beach]]
What we have done in (6) is to divide the predicate into a verb (went, the past tense of go) and a prepositional phrase (to the Beach). And, of course, we can further divide the prepositional phrase into a preposition (to) and its noun phrase object (the Beach).
Our grammatical knowledge also allows us to make judgments about reference; we know that some lexical items refer to others in the same sentence or in the discourse (i.e., extra-sentential reference). For example:
7. Dillon hit himself
8. Dillon hit him
We know that himself must refer to Dillon in (7), yet in (8) we know that him cannot refer to Dillon.
Our syntactic knowledge also allows us to make judgments of "sameness of meaning." For example:
9. Dillon hit Joelle
10. Joelle was hit by DillonWe know that although the focus may be different in (9) and (10), the basic meaning is the same.
Finally, our syntactic knowledge allows us to recognize structural ambiguities, sometimes subtle ones:
11. The mechanic fixed the car in the garage
If some thought is given to what (11) can, in principle, mean, we come to the conclusion that it can mean either something like "it was in the garage where the mechanic fixed the car" or it can mean "it was the car in the garage (instead of the car outside the garage) that the mechanic fixed." These meanings, as we shall see later, are reflected by where the prepositional phrase (in the car) fits into the phrasal geometry of the sentence.
It is likely the reader cannot verbalize the rules and principles that the judgments about examples (1)-(11) were based on, basically because knowledge of language is tacit or unconscious. The job of the theoretical syntactician is to observe, hypothesize, and test what this knowledge consists of.
Lexical and Functional Categories
Categories are theoretical constructs that linguists use to explain the fact that some words behave differently than others. Instead of using vague notions like "nouns are persons, places, or things" (the word run can be a noun or a verb), "verbs refer to actions" (destruction is an "action," but is actually a noun), and "prepositions are words referring to locations" (La Jolla is a location, but is a noun), linguists have looked to phonological, morphological, and distributional evidence to determine or rationalize lexical categorization, or parts of speech. For example, phonologically, the primary stress often falls on the first syllable of multisyllabic nouns (e.g., PERmits, RECords), yet on verbs the primary stress often falls on the second syllable (e.g., perMITS, reCORDS).
Morphologically, nouns can be pluralized (boys, women) and verbs cannot. Nouns and verbs can form complex words made up of more than one morpheme, but prepositions cannot; they are invariant.
Distributionally, nouns occur in particular and in different parts of a sentence than do verbs; thus, they cannot be substituted for each other (they are said to be in complementary distribution). For example, nouns can be pre-modified by adjectives (very big boy, pretty woman, etc.) yet verbs cannot ([*]very big know); nouns can be quantified and specified (e.g., made definite or indefinite) (e.g., the boy, a boy), yet verbs cannot ([*]a / the know). So, a verb cannot be substituted for a noun, and vice versa (Dillon kissed Joelle, [*]Kissed Dillon Joelle).
Indeed, substitution is one constituency test that linguists use to help determine the category of a lexical item.
Because of these phonological, morphological, and distributional facts, linguists have hypothesized a limited set of lexical categories such as Nouns, Verbs, Adjectives, and Prepositions, as well as a set of functional categories like Determiners (the, this, some, many, etc.), Complementizers (that, whether, for, etc.), and Inflections (modals such as will and should, for example, and tense and agreement morphemes). Claiming that there are distinct categories that behave differently allows the linguist to make general statements, like "nouns can be pluralized, verbs cannot." Such general statements allow lexical items to be represented economically. For example, because only nouns can be pluralized, it is not necessary to represent the plural noun separately from its singular version. All that is needed is a representation of the singular noun and a rule that states that nouns can be pluralized; this will automatically generate the plural form for the noun. Importantly, this productive mechanism simplifies the acquisition process that the child undergoes. That is, the child does not need to "memorize" each plural form for each singular noun counterpart; all the child needs to know is the rule for the plural and the fact that any noun can have a plural form (although the child will indeed have to memorize irregular forms). This emphasis on language acquisition forms the basis for much of linguistic theory.
Journal of Speech, Language & Hearing Research, Apr97, Vol. 40 Issue 2, p254, 19p.