After a lecture on cosmology a

After a lecture on cosmology and the structure of the solar system, William James was accosted by a little old lady who told him that his view of the earth rotating round the sun was wrong.

“I’ve got a better theory,” said the little old lady.
“And what is that, madam?” inquired James politely.
“That we live on a crust of earth which is on the back of a giant turtle.”
“If your theory is correct, madam,” he asked, “what does this turtle stand on?”
“You’re a very clever man, Mr. James, and that’s a very good question,” replied the little old lady, “but I have an answer to it. And it’s this: the first turtle stands on the back of a second, far larger, turtle, who stands directly under him.”
“But what does this second turtle stand on?” persisted James patiently.
To this, the little old lady crowed triumphantly, “It’s no use, Mr. James, it’s turtles all the way down.”
Adapted from Ross (1967)
In the last chapter, we moved from the general categories and concepts of traditional grammar to more specific methods of describing the structure of phrases and sentences. When we concentrate on the structure and ordering of components within a sentence, we are studying the syntax of a language. The word “syntax” comes originally from Greek and literally means “a putting together” or “arrangement.” In earlier approaches to the analysis of syntax, as we saw in Chapter 7, there was an attempt to produce an accurate description of the sequence or ordering “arrangement” of elements in the linear structure of the sentence. In more recent attempts to analyze syntactic structure, there has been a greater focus on the underlying rule system that we use to produce or “generate” sentences.
Syntax
When we set out to provide an analysis of the syntax of a language, we try to adhere to the “all and only” criterion. This means that our analysis must account for all the grammatically correct phrases and sentences and only those grammatically correct phrases and sentences in whatever language we are analyzing. In other words, if we write rules for the creation of well-formed structures, we have to check that those rules, when applied logically, won’t also lead to ill-formed structures.
For example, we might say informally that, in English, we put a preposition (near) before a noun (London) to forma prepositional phrase (near London). However, if we use this as a rule of the grammar to create structures, we will end up producing phrases like *near tree or
*with dog. These don’t seem to be grammatically correct, so we mark them with an asterisk *. We clearly need to be more careful in forming this rule. We might have more success with a rule stating that we put a preposition before a noun phrase (not just a noun). In Chapter 7, we saw that a noun phrase can consist of a proper noun (London), a pronoun (you) or a combination of an article (a, the) and a noun (tree, dog), so that the revised rule can produce these well-formed structures: near London, with you, near a tree, with the dog. When we have an effective rule such as “a prepositional phrase in English consists of a preposition followed by a noun phrase,” we can imagine an extremely large number of English phrases that could be produced using this rule. In fact, the potential number is unlimited. This reflects another goal of syntactic analysis, which is to have a small and finite (i.e. limited) set of rules that will be capable of producing a large and potentially infinite (i.e. unlimited) number of well-formed structures. This small and finite set of rules is sometimes described as a generative grammar because it can be used to
“generate” or produce sentence structures and not just describe them.
This type of grammar should also be capable of revealing the basis of two other phenomena: first, how some superficially different sentences are closely related and, second, how some superficially similar sentences are in fact different.
Deep and surface structure
Two superficially different sentences are shown in these examples:
Charlie broke the window.
The window was broken by Charlie.
In traditional grammar, the first is called an active sentence, focusing on what Charlie
did, and the second is a passive sentence, focusing on The window and what happened

to it. The distinction between them is a difference in their surface structure, that is, the different syntactic forms they have as individual English sentences. However, this superficial difference in form disguises the fact that the two sentences are very closely related, even identical, at some less superficial level.
This other “underlying” level, where the basic components (Noun Phrase + Verb + Noun Phrase) shared by the two sentences can be represented, is called their deep structure. The deep structure is an abstract level of structural organization in which all the elements determining structural interpretation are represented. That same deep structure can be the source of many other surface structures such as It was Charlie who broke the window and Was the window broken by Charlie?. In short, the grammar must be capable of showing how a single underlying abstract representation can become different surface structures.
Structural ambiguity
Let’s say we have two distinct deep structures. One expresses the idea that “Annie had an umbrella and she bumped into a man with it.” The other expresses the idea that “Annie bumped into a man and the man happened to be carrying an umbrella.” Now, these two different versions of events can actually be expressed in the same surface structure form: Annie bumped into a man with an umbrella. This sentence provides an example of structural ambiguity. It has two distinct underlying interpretations that have to be represented differently in deep structure.
The comedian Groucho Marx knew how to have fun with structural ambiguity. In the film Animal Crackers, he first says I once shot an elephant in my pajamas, then follows it with How he got into my pajamas I’ll never know. In the non-funny interpretation, part of the underlying structure of the first sentence could be something like: “I shot an elephant (while I was) in my pajamas.” In the other (ho, ho) interpretation, part of the underlying structure would be something like: “I shot an elephant (which was) in my pajamas.” There are two different underlying structures with the same surface structure.
Phrases can also be structurally ambiguous, as in expressions like small boys and girls. The underlying interpretation can be either “small boys and (small) girls” or “small boys and (all) girls.” Our syntactic analysis will have to be capable of showing the structural distinction between these underlying representations.
Recursion
The rules of the grammar will also need the crucial property of recursion. Recursive
(“repeatable any number of times”) rules have the capacity to be applied more than

once in generating a structure. For example, we can have one prepositional phrase describing location (on the table) in the sentence The gun was on the table. We can also repeat this type of phrase, using different words (near the window), for as long as the sentence still makes sense (in the bedroom). So, in order to generate a sentence such as The gun was on the table near the window in the bedroom, we must be able to repeat the rule that creates a prepositional phrase over and over again.
We must also be able to put sentences inside other sentences. For example, when we produce a sentence such as Cathy knew that Mary helped George, we do so with the sentence Mary helped George inside it. And those two sentences can be generated inside another sentence such as John believed that Cathy knew that Mary helped George. In principle, there is no end to the recursion that would produce ever longer versions of complex sentences with this structure.
Basically, the grammar will have to capture the fact that a sentence can have another sentence inside it or that a phrase can be repeated as often as required. We should note that recursion of this type is not only a feature of grammar, but can also be an essential part of a theory of cosmic structure, as in the role of turtles in one little old lady’s view of the universe (in the introductory quotation).
Tree diagrams
One of the most common ways to create a visual representation of syntactic structure is through tree diagrams. We can use the symbols introduced in Chapter 7 (Art = article, N = noun, NP = noun phrase) to label parts of the tree as we try to capture the hierarchical organization of those parts in the underlying structure of phrases and sentences. So, we can take the information in a labeled and bracketed format, shown on the left, and present it in a tree diagram, shown on the right.
Although this kind of “tree,” with its “branches,” shown on the right, seems to grow down rather than up, it functions rather well as a diagram representing all the grammatical information found in the other analysis on the left. It also shows very explicitly that there are different levels in the analysis. That is, there is a level of
analysis at which a constituent such as NP is represented and a different, lower, level at
NP NP

Art N Art N
[The] [girl]
Figure 8.1

The girl

S

NP VP

Art N V NP
Art N

The girl saw a dog
Figure 8.2
which a constituent such as N is represented. This type of hierarchical organization can be illustrated in a tree diagram for a whole sentence, beginning at the top with S.
If we start at the top of the tree diagram, we begin with a sentence (S) and divide it into two constituents (NP and VP). In turn, the NP constituent is divided into two other constituents (Art and N). Finally, one word is selected that fits the label Art (the) and another that fits N (girl). You can go through the same procedure with the VP branches.
Symbols used in syntactic analysis
We have already encountered some symbols that are used as abbreviations for syn- tactic categories. Examples are “S” (= sentence), “NP” (= noun phrase), “N” (= noun), “Art” (= article), “V” (= verb) and “VP” (= verb phrase). Others, such as “PP” (= prepositional phrase), seem fairly transparent. There are three more symbols that are commonly used in syntactic description.
The first is in the form of an arrow ?. It can be interpreted as “consists of” or
“rewrites as.” It is typically used in the following type of rule:
NP ? Art N
This is simply a shorthand way of saying that a noun phrase (NP) such as the dog
consists of or rewrites as (?) an article (Art) the and a noun (N) dog.
The second symbol is a pair of round brackets ( ). Whatever occurs inside these round brackets will be treated as an optional constituent. For example, we can describe something as the dog or the small dog. We can say that both the dog and the small dog are examples of the category noun phrase (NP). When we want to use a noun phrase in
English, we can include an adjective (Adj) such as small, but we don’t have to. It’s an optional constituent in a grammatically well-formed noun phrase. We can represent this observation in the following type of rule:

NP ? Art (Adj) N
This shorthand notation expresses the idea that a noun phrase rewrites as (?) an article (Art) and a noun (N), with the option of including an adjective (Adj) in a specific position between them. We use the round brackets to indicate that the adjective is optional. So, we can use this notation to generate the dog, the small dog, a cat, a big cat, the book, a boring book and an endless number of other similar noun phrases.
The third symbol is in the form of curly brackets { }.These indicate that only one of the elements enclosed within the curly brackets must be selected. We use these types of brackets when we want to indicate that there is a choice from two or more constituents. For example, we have seen already that a noun phrase can consist of an expression such as the dog (article plus noun), or it (pronoun), or Cathy (proper noun). Using the abbreviations “Pro” (for pronoun) and “PN” (for proper noun), we can try to capture this observation about English with three separate rules, as shown on the left. However, it is more succinct to write one rule, as shown in the middle or on the right, using curly brackets and including exactly
the same information.
NP ? Art N

Art N

NP ? Pro NP ?

fPro g

NP ? {Art N, Pro, PN}

NP ? PN PN

It is important to remember that, although there are three constituents inside these curly brackets, only one of them can be selected on any occasion.
The list of common symbols and abbreviations is summarized here.

S sentence NP noun phrase PN proper noun N noun VP verb phrase Adv adverb V verb Adj adjective Prep preposition Art article Pro pronoun PP prepositional phrase
* ungrammatical sentence
? consists of / rewrites as
( ) optional constituent
{ } one and only one of these constituents must be selected
Phrase structure rules
When we use a tree diagram format, we can think of it in two different ways. In one way, we can simply treat it as a static representation of the structure of the sentence shown at the bottom of the diagram. We could then propose that, for every single sentence in English, a tree diagram of this type could be drawn. An alternative view is to treat the tree diagram as a dynamic format, in the sense that it represents a way of generating not only that one sentence, but a very large number of other sentences with similar structures.
This second approach is very appealing because it would enable us to generate a very large number of sentences with what look like a very small number of rules. These rules are called phrase structure rules. As the name suggests, these rules state that the structure of a phrase of a specific type will consist of one or more constituents in a particular order. We can use phrase structure rules to present the information of the tree diagram in another format. That is, the information shown in the tree diagram on
the left can be expressed in the phrase structure rule on the right.
NP

Art N NP Art N
Figure 8.3
According to this rule, “a noun phrase rewrites as an article followed by a noun.” The first rule in the following set of simple (and necessarily incomplete) phrase
structure rules states that “a sentence rewrites as a noun phrase and a verb phrase.” The second rule states that “a noun phrase rewrites as either an article plus an optional adjective plus a noun, or a pronoun, or a proper noun.” The other rules follow a similar pattern.

S ? NP VP
NP ? {Art (Adj) N, Pro, PN} VP ? V NP (PP) (Adv)
PP ? Prep NP
Lexical rules
Phrase structure rules generate structures. In order to turn those structures into recognizable English, we also need lexical rules that specify which words can be

used when we rewrite constituents such as N. The first rule in the following set states that “a proper noun rewrites as Mary or George.” (It’s a very small world.)

PN ? {Mary, George} N ? {girl, dog, boy} Art ? {a, the}
Pro ? {it, you}
V ? {followed, helped, saw}
We can rely on these rules to generate the grammatical sentences shown below as (1)
to (6), but not the ungrammatical sentences shown as (7) to (12).
(1) A dog followed the boy. (7) *Dog followed boy. (2) Mary helped George. (8) *The helped you boy. (3) George saw the dog. (9) *George Mary dog.
(4) The boy helped you. (10) *Helped George the dog. (5) It followed Mary. (11) *You it saw.
(6) You saw it. (12) *Mary George helped.
As a way of visualizing how the phrase structure rules form the basis of these sentences, we can draw the tree diagrams for sentences (1) and (6).
(1) S (6) S
NP VP NP VP
Art N V NP Pro V NP Art N Pro
A dog followed the boy You saw it
Figure 8.4
Movement rules
The very small set of phrase structure rules just described is a sample of what a more complex phrase structure grammar of English, with many more parts, would look like. These rules can be treated as a representation of the underlying or deep structures of

sentences in English. One feature of these underlying structures is that they will generate sentences with a fixed word order. That is convenient for creating declarative forms (You will help Mary), but not for making interrogative forms, as used in ques- tions (Will you help Mary?). In making the question, we move one part of the structure to a different position. This process is based on a movement rule.
In order to talk about this process, we need to expand our phrase structure rules to include an auxiliary verb (Aux) as part of the sentence. This is illustrated in the first rewrite rule below. Auxiliary verbs (sometimes described as “helping” verbs) take different forms in English, but one well-known set can be included in the rudimentary lexical rule for Aux below. We also need a lexical rule that specifies the basic forms of the verbs, shown as the third rewrite rule below.

S ? NP Aux VP
Aux ? {can, could, should, will, would} V ? {follow, help, see}
With these components, we can specify a simple movement rule that is involved in the creation of one basic type of question in English.

NP Aux VP ? Aux NP VP
This type of rule has a special symbol ? and can be illustrated in the process of one tree, on the right, being derived from the tree on the left.
S ? S

NP Aux VP Aux NP VP

Pro V NP Pro V NP PN PN
You will help Mary Will you help Mary
Figure 8.5

Using this simple rule, we can also generate these other questions:

Can you see the dog? Should George follow you? Could the boy see it? Would Mary help George?

These are all surface structure variations of a single underlying structure. However, we still have not incorporated recursion.
Back to recursion
The simple phrase structure rules listed earlier have no recursive elements. Each time we start to create an S, we only create a single S (sentence structure). We actually need to be able to include sentence structures within other sentence structures. In traditional grammar, these “sentence structures” were described as “clauses.” We know, for example, that Mary helped George is a sentence. We can put this sentence inside another sentence beginning Cathy knew that [Mary helped George]. And, being tedi- ously recursive, we can put this sentence inside another sentence beginning John believed that [Cathy knew that [Mary helped George]].
In these sentences, two new proper nouns and two new verbs have been used. We have to expand our earlier set of lexical rules to include PN ? {John, Cathy} and V ?
{believed, knew}. After verbs such as believe and know, as in these examples, the word
that introduces a complement phrase.

Mary helped George.
Cathy knew that Mary helped George.
John believed that Cathy knew that Mary helped George.
Complement phrases
The word that, as used in these examples, is called a complementizer (C). The role of that as a complementizer is to introduce a complement phrase (CP). For example, in the second sentence (Cathy knew …), we can identify one CP which contains that plus Mary helped George. We already know that Mary helped George is a sentence (S). So, we are now in a position to define a CP in the following way: “a complement phrase rewrites as a complementizer and a sentence,” or CP ? C S.
We can also see from the same sentence that the complement phrase (CP)
comes after a verb (V) knew. This means that we are using the CP as part of a verb phrase (VP), as in knew that Mary helped George. So, there must be another rule that says: “a verb phrase rewrites as a verb and complement phrase,” or VP ? V CP.

If we now look at these two new rules in conjunction with an earlier rule, we can see how recursion is built into the grammar.

S ? NP VP VP ? V CP CP ? C S

We begin with S on the left and, as we rewrite symbols, we eventually have S on the right, allowing us to go back to the beginning and go through the set of rules again (and again). This means that we can, in principle, use these rules to create an endless sentence containing other sentence structures. In practice, it allows us to draw the following tree diagram and provide a clear representation of the syntactic structure of
this one fairly complex sentence.

S

NP VP

V CP

C S

NP VP

V CP

C S

NP VP
V NP PN PN PN PN
John believed that Cathy knew that Mary helped George
Figure 8.6

As we try to capture more aspects of the structure of complex English sentences, we inevitably need to identify more rules and concepts involved in the analysis of syntax. (We’ve barely scratched the surface structures.) However, having explored some of the basic issues and methods of syntactic analysis in order to talk about structure in language, we must move on to consider how we might incorporate the analysis of meaning in the study of language.

Study questions
1 What is the “all and only” criterion?
2 Do phrase structure rules represent deep structure or surface structure?
3 In what ways are these expressions structurally ambiguous?

(a) The parents of the bride and groom were waiting outside. (b) We met an English history teacher.
(c) Flying planes can be dangerous.
(d) The students complained to everyone that they couldn’t understand.

4 Which of the following expressions would be generated by this phrase structure rule: NP ? {Art (Adj) N, Pro, PN}?

(a) a lady (c) her (e) the widow (b) the little girl (d) Annie (f) she’s an old woman
5 Which of these sentences would result from applying the rule: NP Aux VP ? Aux
NP VP?

(a) John will follow Mary. (c) Can George see the dog?
(b) You knew that Cathy helped the boy. (d) Should you believe that Mary saw it?

6 Using information from the phrase structure rules presented in this chapter, complete the following tree diagrams.

(a) S

NP VP

(b) S

NP VP

Art

NP
S

VP

The girl saw you

N

Figure 8.7

John knew that you helped the small boy
Tasks
A What is the distinction made between “competence” and “performance” in the study of syntax?
B What is meant by the expression “an embedded structure”? Were there any examples in this chapter?
C Which of the following two tree diagrams could be used to represent the underlying structure of the sentence: George saw the boy with
a telescope?
(i) S (ii) S

NP VP NP VP
V NP PP V NP
Figure 8.8

Art N PP

D In spoken English, the sequence want to is sometimes contracted to wanna, as in I don’t wanna go or What do you wanna do tonight?. However, as illustrated in the following set of sentences, there are some structures where want to cannot be contracted. English-speaking children know how to use wanna in the right places (and none of the wrong places) at a very early age. Can you work out what it is that they know about using wanna?

(1) Who do you want to or wanna visit?
(2) Who would you want to or wanna go out with?
(3) How many of your friends do you want to or wanna invite to the wedding?
(4) Who do you want to (*wanna) win the game?
(5) Who would you want to (*wanna) look after your pets?
(6) How many of your friends do you want to (*wanna) stay with us?

E The following simplified set of phrase structure rules describes some aspects of the syntax of a language called Ewe, spoken in West Africa. Based on these rules, which of the following sentences (1-10) should have an asterisk * before them?

S ? NP VP N ? {oge, ika, amu} NP ? N (Art) Art ? ye
VP ? V NP V ? {xa, vo}

(1) Oge xa ika (6) Vo oge ika
(2) Ye amu vo oge (7) Amu ye vo ika (3) Ika oge xa ye (8) Ye ika xa ye oge (4) Oge ye vo ika ye (9) Xa amu ye
(5) Amu xa oge (10) Oge ye xa amu

F Using these simple phrase structure rules for Scottish Gaelic, identify (with *) the ungrammatical sentences below and draw tree diagrams for the grammatical sentences.

S ? V NP NP NP ? {Art N (Adj), PN} Art ? an N ? {cu, duine, gille} Adj ? {ban, beag, mor} PN ? {Calum, Mairi, Tearlach} V ? {bhuail, chunnaic, fhuair}
(1) Calum chunnaic an gille.
(2) Bhuail an beag cu Tearlach. (3) Bhuail an gille mor an cu. (4) Chunnaic Tearlach an gille. (5) Ban an cu an duine beag. (6) Fhuair Mairi an cu ban.

Discussion topics/projects
I There is a principle of syntax called “structure dependency” that is often used to show that the rules of language structure depend on hierarchical organization
and not on linear position. For example, someone trying to learn English might be tempted to think that questions of the type in (2) are formed simply by
moving the second word in a statement (1) to become the first word of a question (2).
(1) Shaggy is tired. (2) Is Shaggy tired?
You will help him. Will you help him?
Using the sentences in (2)-(6), try to decide if this is the best way to describe how all of these English questions are formed and, if it is not, try to formulate a better rule.

(3) Are the exercises in this book too easy?
(4) Is the cat that is missing called Blackie?
(5) Will the price of the new book you’ve ordered be really expensive?
(6) Was the guy who scored the winning goal in the final playing for love or money?

(For background reading, see chapter 4 of Fromkin et al., 2007.)

II We could propose that passive sentences (George was helped by Mary) are derived from active structures (Mary helped George) via a movement rule such as the following:

(active ) NP1 V NP2 ? NP2 be V-ed by NP1 (passive)

Note that the tense, past or present, of the V (e.g. helped) in the active structure determines the tense of be in the passive structure (e.g. was helped).
Which of the following active sentences can be restructured into passive sentences
using this rule? What prevents the rule from working in the other cases?

(1) The dog chased the cat.
(2) Snow White kissed Grumpy. (3) He loves them.
(4) Betsy borrowed some money from Christopher. (5) The team played badly.
(6) The bank manager laughed. (7) They have two children.
(8) The duckling became a swan.
(9) Someone mentioned that you played basketball. (10) The police will arrest violent demonstrators.

(For background reading, see chapter 5 of Morenberg, 2003).

Further reading
Basic treatments
Miller, J. (2008) An Introduction to English Syntax (2nd edition) Edinburgh University Press
Thomas, L. (1993) Beginning Syntax Blackwell
More detailed treatments
Morenberg, M. (2003) Doing Grammar (3rd edition) Oxford University Press Tallerman, M. (2005) Understanding Syntax (2nd edition) Hodder Arnold Specifically on English syntax
Radford, A. (2004) English Syntax Cambridge University Press
On generative grammar
Baker, M. (2001) The Atoms of Language: The Mind’s Hidden Rules of Grammar Basic Books
On structural ambiguity
Pinker, S. (1994) The Language Instinct (chapter 4) William Morrow
Tree diagrams
Carnie, A. (2002) Syntax (chapter 12) Blackwell
On Gaelic syntax
Brown, K. and J. Miller (1991) Syntax: A Linguistic Introduction to Sentence Structure (2nd edition) Routledge
Other references
Fromkin, V., R. Rodman and N. Hyams (2007) An Introduction to Language (8th edition) Thomson