Theta Theory in English and French

By Badr Assila,

badrassila2005[at]hotmail.com

Get the List of 4,500+ Translation Agencies Now!
No Recurring Membership Fees!

As the title of this paper indicates, I am going to focus on Theta Theory in English and French. In this way, the first chapter will be concerned with Theta Theory in English. In the first section, I am going to discus Universal Grammar and its sub-theories, such as X-bar theory, Case theory, Government theory, Binding theory, Bounding theory, Control theory and a brief introduction to Theta theory. The second section will be devoted to Theta Theory, trying to present it from different points of view, namely that of Riemsdijk, Haegman, Radford and Sadiqi. In the third section, I will present the different types of Theta roles. Thereafter, I will provide the two major principles on which Theta Theory bases its assignment of Theta roles, namely the Theta-Criterion and the Projection Principle. Section five of this chapter is going to define Case Theory and its role in Theta role assignment. Finally, I am going to conclude this chapter with a discussion of Theta Theory and how it provides a good argument for movement processes in English, namely NP movement and Wh movement.

The second chapter of this Monograph will discuss Theta Theory in French. In the first section, I will dwell on the different types of Theta roles acknowledged in French. Next, I am going to deal with Theta role assignment and the two major principles of Theta Theory, namely the Theta Criterion and the Projection Principle. In the third section, I will concentrate on how Case theory interferes in Theta role assignment. Finally, to conclude my chapter, I will discuss the interaction between Theta Theory and movement processes, namely NP movement and Wh movement.

I-1- The Organization Of Universal Grammar:

According to Trask (1993), the term “Universal Grammar” [UG] refers to “ all the grammatical properties, which hold for the grammars of all the existing and possible languages, which define the notion ‘possible grammars’”. The elucidation of UG is one of the chief goals of modern linguistic theory.

In other terms, Chomsky presents the concept of UG- or alternatively, the “initial state”- as a group of principles that constitute the child’s first knowledge of the structure of human grammars. It is also a set of principles that allow for the development of a particular grammar through contact with a particular linguistic environment. This theory of UG consists of many sub theories that I may list as follows:

X-Bar Theory:

According to Sadiqi (1992), X-bar theory represents the information of phrase structure rules with a new version, since it introduces the notion of intermediate projection, namely X’.

In other terms, X-bar theory presents a new approach to phrase structure, making use of the following principles:

XP → Spec ; X’.
X’ → X°; Compl.

According to these principles, a lexical head is intermediary projected into an X’ which in turn maximally projected into an XP, as indicated by the following example:

(1) a- Students of physics.

-b-

According to Haegman (1991), X-bar theory brings out what is common in the structure of phrases. For example, all phrases are headed by a lexical head which is a zero projection (X°); this theory distinguishes between three types of projection: complements combine with X to form an X’ projection, adjuncts combine with X’ to create a copy of X’, and the specifier combines with the topmost X’ to form the maximal projection XP. X-bar principles are assumed to be universal, but the order of constituents with respect to the head of the projection is not universally fixed.

Case Theory:

According to Trask (1993), Case Theory is one of the principal modules in GB, consisting of various case-marking conventions and the case filter. This theory is responsible for insuring that every overt NP in a sentence is marked as possessing a case, such as nominative, accusative, or genitive, as required by the case filter. Case Theory recognizes two types of case assignment, namely “inherent” case, which is assigned at D-structure, and “structural” case, which is assigned at S-structure.

According to Radford (1988), English has three different case-forms illustrated by the paradigm below:

Control Theory:

According to Trask (Ibid), this theory is a module in UG, which deals with the phenomenon by which a VP complement with no overt subject is interpreted semantically as having some NP as subject, either overtly in the sentence or arbitrary (unspecified). In this theory, a non-overt subject is conventially represented by a distinct empty category, called PRO. Subject control is when the PRO is controlled by the subject, while object control is when the PRO is controlled by the object. These two types of control are known as “functional control”. In sentences such as “[PRO smoking] causes cancer], PRO exhibits arbitrary control, and this type is known as “arbitrary control”.

Trace Theory:

According to Trask (Ibid), Trace Theory studies the empty category left behind in a particular location by the movement of some element out of that position. The use of traces allows a tree to ‘remember’ earlier stages of a derivation, and traces can be regarded as a formalization of certain aspects of the earlier derivational constraints. GB recognizes two types of traces: NP-trace and WH-trace. Trace theory and movement theory are interconnected, since after movement theory applies a trace is left behind.

Let’s consider the following example:

(2) a- Will john go to school?

b- [et] john will go to school?

This sentence has undergone a type of movement known as “I-to-C-Movement”, which can be presented in the following tree diagram.

c-

Binding Theory:

This theory deals with most co-reference phenomena among NP’s, including empty categories. Both overt NP’s and empty categories are divided into the following types, using the two binary features [anaphoric] and [pronominal]:

According to Chomsky (1981), these classes are subject to the following binding principles:

1- An anaphor [+a] is bound in its governing category
2- A pronominal [+p] is free in its governing category
3- An R-expression is free everywhere.

According to this theory, a constituent binds another constituent if and only if it satisfies the following Binding Condition:

Let’s deal with this condition in the light of the tree diagram in (3) below:

According to the above condition, the NP {John} and the NP {himself} are coindexed, and the NP {John} c-commands the NP {himself}. Therefore, we can say that the NP {john} binds the NP {himself}.

Government Theory:

According to Trask (1993), this theory is one of the modules recognized in GB. Its function is to ensure that certain types of structural relations hold between nodes in trees; its principal requirement is the Empty Category Principle {ECP}.

According to Aoun and Sportiche (1983), Government can be formulated as follows:

Let’s deal with this condition through the following tree diagram:

(4)

In the above structure, V governs NP¹; it does not govern NP², which is protected by the maximal projection PP.

Bounding Theory:

This theory deals with constraints upon the permissible degree of separation in a structure of two elements that are both involved in the statement of a single grammatical unit. This theory bears an important principle, which is called “the Subjacency Condition”:

This condition is illustrated by the following sentence, derived as indicated in (5-b):

(5) a- The fur seems to be certain to fly.

In (5-b), the NP [the fur] has moved to a near empty position, which is
Marked [NPe²]; then, to make this sentence grammatical, the NP [the fur] has to move to the empty position marked by [NPe¹].

Movement Theory:

In GB, movement theory is a theory, which describes the movement or displacement of a constituent from one position to another. This movement is controlled by rules that relate the two sequential levels of structure, i.e. S-structure and D-structure. Only a single very simple and general movement rule is posited, namely the rule of Alpha Movement. In Movement Theory, many instances of movement are recognized, especially NP-movement and WH-movement.

Let’s consider the following example:

(6)-a- Lisa seems to be happy.

In (6-b), the NP [Lisa] has undergone movement to occupy an empty position marked by [e].

Theta Theory:

θ-theory is the module that deals with the valency requirements of verbs. It incorporates a set of participant roles, called Theta Roles. Their distribution in sentence structure is mediated chiefly by the Projection Principle and the Theta-Criterion. This theory is going to be presented in the coming sections.

I-2- On Theta Theory:

Thematic theory, or theta theory, is a sub-theory of universal Grammar, which deals with the valency requirements of verbs. As I mentioned in the previous section, it incorporates a set of principles regulating the assignment of thematic roles. In this way, Riemsdijk (1986) defines theta theory as the basic logical notion “argument of”, a notion that any theory of Grammar must account for. He goes on to add that the aim of this theory is to determine which NP can be an argument of a verb. That’s why to designate arguments of a verb, terms as agent, goal, patient …are commonly used.

In the same way, Sadiqi (1992) confirms that theta theory is to determine the semantic relationship between constituents in a structure. To present these relationships, this theory employs two basic principles, namely the theta criterion and the projection principle.

Let’s consider the following example:

(7) John killed Mary.

According to Theta theory, the above structure is built up on certain semantic relationships. Thus, the NP [john] bears the theta role “agent” and the NP [Mary] receives the role “patient”.

According to Haegman (1991), θ-theory accounts for the semantic relationships between a verb and its arguments through the assignment of theta roles. That a verb theta marks its arguments means that it assigns theta roles to these arguments. This assignment is controlled by two major principles, namely the theta criterion and the projection principle.

According to Radford (1988), theta theory is a module in universal Grammar, which claims that there are certain thematic relations that relate arguments in a structure. According to this theory, each argument receives a semantic role also known as theta role. The distribution of these roles is determined by two major principles, i.e. the theta criterion and the projection principle. (c.f. Sect. I-4)

The following example illustrates the above claim:

(8) John gave Mary a flower.

In this example, we have three NP’s that are related to each other on the basis of abstract thematic relationships. These relationships are established through the assignment of theta roles. In this sentence, every NP receives a specific θ-role. Therefore, [john] is assigned the role ‘agent’, [Mary] received the role ‘goal’ while the third NP [a flower] has the theta role ‘patient’.

I-3- θ-Roles: Definition & Typology

I-3-1- Definition:

According to Trask (1993), “θ-role” is the usual term for one of the semantic roles recognized in GB and assigned by predicates to their arguments by the requirements of Theta Theory. These roles are necessary to represent the argument structure of the verb; every predicate comes with a predefined set of θ-roles, which it requires to be expressed if the sentence is to be grammatical.

According to Jackendoff (1990), Theta Roles are relational notions defined structurally over conceptual structure, with a status precisely comparable to that of the notions subject and object in many syntactic theories. They are not marked as annotations to D-structure or to predicate argument structure nor are they specified at the special level of representation.

I-3-2- Typology of theta roles:

A Theta Role represents the semantic relationship of arguments with the predicate they are arguments of. These roles are of many types, which I may list as follows:

I-4- The Theta-Criterion & the projection Principle:

As mentioned in previous sections, Theta Theory operates at D-structure through the assignment of θ-roles to arguments. This theory is based on two fundamental principles, namely the Theta-Criterion and the Projection Principle. In this section, I will dwell on each of these.

I-4-1- The Theta-Criterion:

According to Chomsky (1981), the Theta-Criterion is to be defined as the following:

This principle means that there must be a one-to-one correspondence between arguments and their Theta- roles.

Let’s consider the following example:

(9)- John killed the bird.

Here the NP [John] is assigned the Theta- role “agent”, while the NP [the bird] is assigned the θ-role “patient”.

Accordingly, an argument that takes two Theta-roles or a θ-role, which is assigned to two arguments, would violate the Theta-Criterion; and therefore, the sentence would be ungrammatical.

The following example illustrates the above claim:

(10)-The key opened by john.

In this example, both the phrases [the key] and [by john] receive the same theta role “agent”, which constitutes a violation of the Theta-Criterion.

Let’s take another example:

(11)-* the navy sank the enemy ship by a torpedo.

According to this sentence, both [the navy] and [by a torpedo] bear the same theta role ‘agent’ which violates the theta criterion. Therefore the sentence is considered to be ungrammatical.

I-4-2- the projection principle:

This principle is one of the major principles of Theta Theory. It can be formulated, as cited by Chomsky (1981), as follows:

The above principle helps us to judge the grammaticality of some syntactic structures.

Let’s take the following rule:

(12)- X’ = X YP

If we replace the variable X by verb and Y by noun, we will get the following structures:

(13)-a- V’ = V NP
(13)-b- V’ = V NP NP
(13)-C-* V’ = V NP NP NP

Thus, according to the Projection Principle, we are able to say that the third structure is ungrammatical, since we have no verb in English that takes three NP complements.

Looking at the Projection Principle cited above, we find that it is too restrictive, since it contains only Subcategorization properties of items; if we implicitly include the Theta-Criterion, as Radford (1988) claims, we might then revise the Projection Principle as follows:

Here, we might assume that the term “lexical” includes both Subcategorization properties as well as thematic ones. Thus, according to the generalized Projection Principle, a verb that takes an agent subject cannot take a theme or a goal subject, while verbs that have a non-thematic subject cannot be inserted in a structure where its subject has received a θ-role.

Let’s take the following examples:

(14) a- John murdered Mary.
(14) b- It seems to be a day of troubles.

In example (14-a), “john” is assigned the θ-role ‘agent’ and “to murder” cannot take a theme subject, while in example (14-b) the verb “seem” has a non-thematic subject and therefore cannot take a thematic subject.

I-5- Theta Theory & Case Theory:

I-5-1- Case Theory:

According to Trask (1993), Case Theory is responsible for ensuring that every overt NP in a sentence should be assigned a case, such as, nominative, accusative, or genitive, as required by the Case Filter¹.

In the same way, Chomsky (1981) supposes that the fundamental properties of Case assignment are as follows:

(15)-a- NP is nominative if governed by AGR
-b- NP is objective if governed by V
-c- NP is oblique if governed by P
-d- NP is genitive if it occurs in the context [NP—X’]
-e- NP is inherently case marked as determined by properties of its [-N] governor.

Case Theory recognizes two types of case assignment, namely “structural” case as in (15-a…-d), which is assigned at S-structure; while the other type is “inherent”, as in (15-e), which is assigned at D-structure.

According to Emonds (1985), the Case Filter constitutes the most general device by which every NP at S-structure must have a case feature. Even though empty NP’s do not have case, they are coindexed with NP’s that do have one. The Case Filter is used to show the well-formedness of S-structure representations.

According to this principle, every NP at S-structure must have a case feature, even if there are instances when empty NPs do not have such a feature, but are coindexed with NPs that are assigned case. Therefore, only empty categories may escape the Case Filter and appear with no case.

Let’s consider the following example:

(16) Peter offered Mary a flower.

In the above example, there are three NPs; the NP [Peter] is assigned nominative case from the AGR element in INFL because it is governed by it. The NP [Mary] is governed and assigned accusative case by the verb [offered]. However, the NP [a flower] cannot be assigned structural case under government given its non-adjacency to the verb [offered]. This NP can only be assigned inherent case at D-structure where the ditransitive nature of the VP is encoded in its sub-categorization frame.

I-5-2- Case Theory & Theta Role Assignment:

According to Sadiqi (1992), there is a parallel between case assignment and θ-role assignment; the head of a phrase that is a V, N, P assign case and theta-role to the elements which depend on them. The AGR element of INFL assigns nominative case to the subject of a sentence.

According to Chomsky (1981), “a verbal element assigns case to an NP that it governs if and only if it assigns a theta role to its subject.”

To clarify this claim, let’s consider the following example:

(17)- John burnt Mary.

In this example, the verb ‘burnt’ assigns an accusative case to its NP object “Mary”, since this NP is under the government of the verb “burnt”. At the same time, the VP “burnt Mary” assigns the theta role ‘agent’ to its NP subject “john”.

There are instances where there is an interaction between case assignment and theta role assignment, although the opposition is found in the passive.

Let’s consider the following example:

Raising verbs like “seem” do not assign case or θ-role to the NP following them. Thus [Peter] moves to the empty subject position marked [e] to receive case from AGR in INFL. In the same way, a verb in the past participle does not assign case to the NP following it. Therefore, the sentence (19-a) is ungrammatical, since the NP [peter] is not assigned case. However, the verb in (19-a) does not assign case to the object, since if we passivise this sentence, we will get a well-formed structure as in (19-b).

(19)-a-* It was hurt Peter
(19)-b- Peter was hurt.

According to Haegman (1991), the Visibility Condition constitutes another element where theta theory and case theory interact. This condition states that “a predicate can only assign a θ-role to NP’s that are visible, and that only abstract case make NP’s visible” Hence, sentences in which argument NP’s without case violate the theta criterion.

Let’s take the following example:

(20)- John travelled to London.

According to the Visibility Condition, the NP [john] is visible and therefore is assigned a theta role “agent”.

According to the Visibility Condition, invisible NP’s cannot be assigned a theta role. Hence, in a sentence like “[PRO] smoking causes cancer”. PRO here receives the so-called ‘null case’, but since PRO here does not receive a theta role, the theta criterion will be violated.

I-6- Theta Theory & Movement Theory:

Movement theory is a theory, which describes the displacement of a constituent from one position to another. This process plays an important role in deriving S-structure from D-structure by displacing elements.

Accordingly, movement theory deals with the displacement of a constituent from one position to occupy another empty position. This movement is restricted by various principles of UG, among which there are the principles of theta theory.

In this way, when we say that a constituent moves, we have to ensure that it preserves all its features, among which the theta role feature.

Let’s consider the following example:

(21) a- [e] seems [john to be friendly]
b- John seems to be friendly.

According to the above example, sentence (21-a-) is ungrammatical, since it violates the Subcategorization requirement of the verb “seem”. Then, when we get the sentence (21-b), we say that the NP [john] has moved to occupy the subject position; this movement has taken place from a θ-position to a θ’-position i.e. a position where no theta role is assigned.

According to the Projection Principle, which is one of the pillars of Theta Theory, NP-movement must take place from one position where a theta role is assigned, i.e. a θ-position, to a position where no theta role is assigned, i.e. a θ’-position.

This is illustrated by the following example:

(22)-a- [e] wants [peter to seem that john is asleep].
(22)-b-* Petter wants [[t] to seem that john is asleep]

The reason for the ungrammaticality of (22-b) is that the NP [Peter] receives no Theta role at D-structure.

According to the theta Criterion, an argument receives only one θ-role, which is in turn assigned to only one argument. Therefore, a constituent moves from one θ-position to a θ’-position so as not to violate this principle.

Let’s take the following example:

In the above sentence, the NP [john] has undergone movement to an empty position where no theta role is assigned. In this movement, the NP [john] is in a θ’ position with no theta role assigned, while its trace bears the theta role “experiencer”, which is transmitted to the NP [john].

Let’s consider another example:

(24) John was read a novel

In this sentence, the NP [a novel] is in a θ-position and cannot move to occupy the subject position, which is filled by the NP [john]. Therefore, the sentence is going to be ungrammatical because it constitutes a clear violation of the theta criterion.

After our brief discussion of Theta theory in English, the aim of this chapter is to examine the same theory but in French. We will show that French supports the existence of theta theory, its elements and principles. In the light of the first chapter, we shall restrict our discussion to the different types of θ-roles acknowledged by French (section one), the assignment of theta roles and principles of theta theory (section two), Case theory and theta role assignment (section three). We will conclude this chapter with a discussion of how Theta theory interacts with movement in French (section four).

II-1- θ-Roles in French:

Having defined theta roles in the first chapter as semantic roles attributed by predicates to their arguments. We are now going to introduce the different theta roles recognized in French.

Types of Theta Role:

II-2- Theta Role Assignment and Principle of Theta Theory:

In this section, we will discuss the assignment of θ- roles in French and the principles of Theta theory, namely the Theta Criterion and the Projection Principle.

II-2-1- The Theta Criterion:

As presented in the first chapter, the Theta Criterion is defined by Chomsky (1981) as follows:

Let’s consider the following example:

(25) Fred a tué sa femme avec un poignard
Fred has killed his wife with a knife

Here, there are three arguments that must be assigned different theta roles. Therefore, NP [Fred] bears the theta role “agent”, the NP [sa femme] receives the theta role “patient”, while the third argument, which is the PP [avec un poignard], is assigned the θ-role “instrument”.

However, if we assign one theta role to two arguments or two theta roles to one argument, we will then violate the Theta Criterion, and therefore the sentence is going to be ungrammatical.

Let’s take the following example:

(26) La clé a fermé la porte par Jean.
The key closed the door by Jean.

This sentence is considered to be ungrammatical, because the two phrases [la clé] and [par Jean] are assigned the same theta role “agent”, and this constitutes a violation to the Theta Criterion.

II-2-2- the Projection Principle:

According to Radford (1988), the Projection Principle is an important principle of Theta theory, which can be presented as follows:

According to this principle, and as presented in the first chapter, the following construction is ungrammatical because no verb in French subcategorises three NP complements.

(27) V’ → V NP NP NP

Also, the Projection Principle shows that a verb that takes an agent subject cannot take an instrument subject, or if a verb takes a non-thematic subject, it cannot have a thematic subject.

Let’s consider the following examples:

(28) Pierre lit le livre
Pierre reads the book

(29) Il semble être un jour de conflit
It seems to be a day of troubles

According to these two examples, “Pierre” is assigned the theta role “agent”, and the verb cannot take a goal subject. In the same way, the verb “semble” takes a non-thematic subject and therefore cannot take a thematic subject.

II-3- Case Theory and θ-Role Assignment:

As discussed in the first chapter, Case theory plays an important role in theta theory and theta role assignment. This theory regulates the assignment of theta roles.

According to Chomsky (1981), “a verbal element assigns case to an NP that it governs if it assigns a theta role to its subject”

The following example clarifies the above claim

(30) Lisa a cassé le vase
Lisa has broken the urn

In this example, the verb “cassé” assigns a theta role “agent” to its subject and an objective case to the NP that it governs.

Let’s take another example:

In this example, the verb “sembler” is a raising verb, which by nature does not assign neither case nor theta role to the NP following it. Therefore, “Pierre” has to move to the empty subject position marked [NPe] to receive case from AGR in INFL.

Another element of interaction between Case theory and Theta theory is, as proposed by Haegman (1991), the Visibility Condition, which states that only visible NP’s are assigned theta roles, and that only abstract case make NP’s visible.

Let’s take the following example:

(32) Mary voyages a Paris
Mary travels to Paris

According to the visibility condition, the NP [Mary] is visible and therefore can be assigned the theta role ‘agent’. This visibility is due to the nominative case attached to the NP [Mary].

Let’s take another example:

(33) [PRO] jouer du sport sera dans votre bénéfice
[PRO] play sport will be good for you

Here, [PRO] receives the so-called “null case” and still cannot be assigned theta role, and then, it will constitute a clear violation of the Theta Criterion.

II-4- Theta Theory and Movement Theory:

In this section, we will present how Movement theory is regulated by principles of Theta theory, mainly the Projection Principle and the Theta Criterion.

According to the Projection Principle, a movement from a θ’-position to θ-position is impossible.

Let’s take the following example:

(34) a- [NPe] veut [Pierre sembler que Mary dort]
b- * Pierre veut [[t] sembler que Mary dort]
Pierre wants [[t] to seem that Mary is sleeping]

The reason why this sentence is ungrammatical is that the NP [Pierre] would receive no theta role at D-structure.

According to the theta criterion, an NP moves from a θ-position to a θ’-position, so as not to violate this principle.

To illustrate, let’s consider the following example:

(35) Pierre semble [t] avoir faim
Pierre seems [t] to be hungry

According to this example, the NP [Pierre] has moved from a θ-position and occupied a θ’-position, leaving a trace behind it. This trace bears the theta role “experiencer” which is transmitted to its antecedent [Pierre]

However, if an NP moves from a θ-position to a θ-position, the Theta Criterion will be violated and the sentence will be ungrammatical.

Let’s take the following example:

(36) Pierre a été lu le livre
Pierre was read the book

In this sentence, the NP [le livre] is originally in a θ-position and cannot move to occupy the subject position, which is filled by the NP [Pierre]. Therefore, the sentence is going to be ungrammatical, as it constitutes a violation of the Theta Criterion.

Let’s take the following example:

(37) Qu’est ce qu’il a frappé ?
What he did beat?
What did he beat?

This sentence has been derived from the following D-structure that has (b-) as its tree representation.

(38) a- [e] il a frappe qu est ce que?

Menu

Free Newsletter