21 Predicate and Argument Structures

Dr. Shamim Fatima

Objectives

The objective of this module is to introduce the key concepts of θ-theory and to show how its emergence is relevant to sentence structure. After the completion of this module the students would be able to understand the thematic structure of English verbs and the proper assignment of theta roles to their arguments which is essential for the formation of a well formed sentence.

Introduction

In this module we discuss one of the important principles of grammar i.e. Theta theory and the argument structure of verbs with special reference to English. Theta theory accounts for the relationship between predicates and their arguments in a systematic manner. Predicates are verbs (e.g. to sleep, to go etc) and the arguments are those noun phrases or prepositional/postpositional

phrases which take part in the completion of a sentence. Theta theory helps us in understanding the nature and roles of a verb and its argument(s). Thematic roles are essential for the formation of a syntactically and semantically correct sentence. It is the verb which decides the number and nature of the arguments in a sentence and assigns ordered thematic roles to them. Theta theory and its components put constraints on the assignment of the role(s) to them.

Theoretical Background

Verbs have always been of the greatest importance always. The structure of sentence is decided by verbs. They can be called as nuclei other than objects of the sentence at which the existence of a sentence depends. In traditional grammar, verbs are classified into two subcategories i.e. intransitive verbs and transitive verbs. Intransitive verbs do not need object to complete a sentence (e.g., to sleep) and transitive verbs need one or more object(s) to complete a sentence (e.g., to eat). Transitive verbs are further put into two categories i.e. verbs which take one object i.e. direct object and the verbs which take two objects i.e. direct object and indirect object. The latter is called ditransitive verbs. Consider following examples:

Such definitions and classifications of verbs in traditional grammar are fit for the verbs which do not take object (intransitive) or take only one object (transitive). However they cannot account for entities which have a different type or relation with the verb. In other words such classifications are not adequate to discuss the nature of all verbs. To clarify, we may look at the following examples:

In (4b-6b), the third element is ‘on the desk’, ‘in the bucket’ and ‘from Chicago’ respectively but they are not objects. They represent location and specification of the objects i.e. book, water and book respectively. Recent developments in the area answer the nature of such elements in regard to their relation with them. They ask the following questions:

If there is no scope to discuss such expressions (e.g. prepositional phrase) in traditional grammar, how can we accommodate in it?
If a given noun phrase is neither ‘Subject’ nor ‘Object’ how can we open our door to such entities?
How can we show what roles such entities play while accounting for an adequate explanation for them?

As there is no scope for such a discussion in traditional grammar, we have to look for an alternative. The Theta Theory has been developed out of result.

Theta Theory

Theta theory is a principle of grammar which deals with the relationship between predicates and arguments. Gruber (1976, (originally 1965)) has first talked about the verbs and their arguments. He said that every argument has a definite kind of relation with the verb in a sentence and these relations and roles are decided by verbs. Thus he introduced the concept of the θ-roles such as ‘agent’, ‘theme’, ‘instrument’, ‘experiencer’, ‘accompaniment’, ‘location’, ‘goal’, ‘source’ etc.

Chomsky (1981) included the theta theory as one of the “subsystems of principles” which is “concerned with the assignment of thematic roles (θ-roles).

Points relevant to Theta theory

Noun phrase (NP) and Postpositional (PoP)/Prepositional phrases (PP) are arguments.
Names, variables, anaphors and pronouns are NP arguments.
Theta roles are the relations which arguments bear with their verbs.
Each and every NP/PP receives a specific theta role from a verb in the sentence.
Idiom chunks or elements (e.g., expletives it or there) inserted to occupy an obligatory position of syntactic structure are not included in NP arguments.
The overt anaphors, R-expressions and pronominals (including empty elements such as
PRO and pro) are all arguments because they have theta roles.
Argument position or A-position- It is a position to which a theta role is assigned. The argument position is termed as a theta position. Object position is always a theta position but the subject position may or may not be a theta position.
A’-position (A-bar position) is a position where no theta role is assigned to an NP.
If the subject position is filled by expletive ‘it’ or ‘there’ then it is not a theta position. If the subject position is filled by an argument, it is a theta position. For instance:

7. The boy is playing football.

8. The child is laughing.

9. It is raining.

10. There is a boy playing in the garden.

In (7), Subject NP the boy and object NP football are arguments. Subject NP the boy is the initiator of the action and the object NP football is the object to which the action is done. In (8), Subject NP child is an argument which is doing action of laughing. In (9) and (10), expletive it and there are at the subject positions respectively. Expletive it and there do not have any semantic relation with verb. In other words, we cannot say what type of role they are playing in sentences. Hence they are not the theta position.

Theta Roles

Each argument has a role to play in a sentence. These roles are decided by their relation with their verbs. The names of these roles are called theta roles (θ-roles). There are different kinds of θ-roles which arguments may have in sentences. Some θ-roles are as follows:

Theta Roles	Definition	Example
Agent	Initiator of the action. It voluntarily does the action.	Raj is playing football.
Patient	Living entity that is the target of the action. In other words, the individual that is affected by an action.	The boy hits Raj.
Theme	Non living entity that undergoes an action which may result in a change of state.	Rice cooked.
Experiencer	Living entity which indicates experiences of involuntary event; the event may be mental or physical.	The child is sneezing.
Goal	Towards which an individual or an entity moves.	The boy is going to the market.
Source	The entity/point from which movement starts.	The train is coming from Delhi
Location	The place where action takes place.	The boy is sleeping in the room.
Instrument	The entity with the help of which the action is done.	The boy is cutting fruits with knife.
Recipient/ Beneficiary	The entity which receives/gets something from someone.	The boy gave a book to Raj.

Note: The types of θ-roles are not fixed. They vary among linguists/scholars. Some scholars do not differentiate between theme and patient. I have distinguished them on the basis of their properties either being animate or inanimate.

Assignment of Theta Roles

The lexical properties of the head of the head phrases determine the assignment of theta roles to their arguments. The verb is the head governor in a sentence which takes NP or PP as its argument(s). Government is a more “local” variety of command (Chomsky 1981a, 1986a, Rizzi 1990) which means to say that the verbs and their arguments must be in local relationship for the assignment of theta roles. As Government theory says that:

“A governs B if A c- commands B and there is no barrier for B c-command by A.”

The assignment of theta roles needs local relation between the verb and its arguments because the verb may assign theta roles to NPs only within its maximal projection i.e. VP. More clearly, the subject argument of the sentence is also within Verb phrase (VP) to take its theta role from the verb. However, Subject i.e [spec, IP] may or may not be in a theta position. It depends on the lexical choices. For example:

11 a. The workers(x) seems [(tx) to protest against the chairman].

b. It seems that the workers will protest against the chairman.

(11a) is a complex sentence which has two verbs i.e. seem and protest. In (11a), the subject of both the verb protest and seem is ‘the workers’. The subject the workers moves to the subject position of seem after taking agent theta role from the verb protest and it leaves a trace (tx) behind it. The verb seem cannot assign any theta role to it. However it forms a chain with the trace (tx) it has left behind in the subject position of the verb protest. Thereby the subject the workers and trace (tx) have the same theta role namely the agent. In (11b) the workers, the subject of the embedded clause, is assigned the agent theta role by ‘protest’. As the subject of the main clause is a non- theta position in (11b) and is occupied by the expletive ‘it’.

VP-Internal Subject Hypothesis

According to this hypothesis, the subject originates in [Spec VP] and after the assignment of theta roles it moves to [Spec IP] position to fulfill another requirement (i.e. for taking case). It is clear from the following tree diagram of (9).

In the tree diagram of (12), Spec NP the boy is within the VP shell and is assigned theta role of agent, NP sweets is assigned the theta role of theme and PP from the shop is assigned the theta role of source.

Theta Criterion (θ-criterion)

Theta criterion is an important component of theta theory which puts constraints on the assignment of theta roles to the verbs. The conditions of theta criterion must be fulfilled in order to generate a correct sentence. Now question is what θ-criterion is. Chomsky defined θ-criterion as follows:

“Each argument bears one and only one theta-role, and each theta role is assigned to one and only one argument.”

It means that the number of θ-roles and the number of arguments must be same. If we will have either the more arguments than the number of θ-roles decided by verb or the less number of arguments than the θ-role(s) decided by verbs, in both cases the sentences get crashed. One to one correspondence between theta roles and their arguments in a sentence is essential for the sentence to be correct. Consider the following example:

13a. Thematic structure of the verb ‘to eat’, <Agent, Theme>

Raj eats pizza.
*Raj eats pizza, burger.

14a. Thematic structure of the verb ‘to kill’, <Agent, Patient>

Raj kills wolf.
*Raj kills.

In (13a), the verb eat needs only two essential theta roles i.e., agent and theme. (13c) has more arguments than the number of theta roles to be assigned. In (14a), the verb to kill has two theta roles i.e. agent and patient to assign but the (14c) has the less number of arguments than it is required to assign theta roles. One theta role (theme) left unassigned due to the lack of the arguments. In both the cases (13c and 14c), the theta criterion gets violated. (13c) can be correct in a condition when the NP arguments pizza and burger are conjoined by the conjunction and. After conjoining they become one NP i.e. pizza and burger and can get one theta role i.e. theme as is clear from (15).

Types of Arguments

Arguments are also classified into categories as per their requirement by verbs. Not every argument is required by every verb. It is the nature of the verbs which decides number and types of arguments verbs may take. According to this requirement, the arguments are of two types:

Essential Arguments
Non-essential Arguments

Essential Argument: The argument without which a sentence will be incomplete is called essential arguments. For instance:

16a. Thematic Structure of ‘to fill’: <Agent, Theme, Location>

16b. Raj is filling water in the bucket

16c. *Raj is filling water

In (16a), we can see that the verb to fill needs three arguments to assign θ-roles of agent, theme and location. (16b) is a well sentence accordingly but (16c) is not correct as one argument for theta role of location is missing there.

Non-essential Argument: The argument which is not required for the sentence to be well formed and correct is called non-essential argument. It provides additional information. Hence it is treated as an adjunct rather than complement. It will be no harm to the grammaticality and meaningfulness of the sentence if we drop it. For instance:

17a. Thematic Structure of the Verb ‘to give’: <Agent, Goal, Theme>

17b. Thematic Structure of the Verb ‘to give’ in (17c): <Agent, Goal, Theme, Source>

17c. Raj gave me a book in the classroom

17d. Raj gave me a book.

In (17a), the verb requires three arguments to assign the θ-roles of agent, goal and theme but (17c) has four arguments i.e. agent ‘Raj’, goal ‘me’, theme ‘a book’, source ‘in the classroom’. It means that one argument is extra there giving additional information i.e. source ‘in the classroom’. The sentence will be complete and correct in case we drop it, as it is clear from (17d).

Classification of Verbs

Verbs are also classified on the basis of the required number and nature of θ-roles their argument(s) might have in their theta grid to generate a grammatically and semantically correct sentence. Intransitive verbs need only one essential argument. There are transitive verbs which need only two arguments while some transitive verbs need three essential arguments for the sentence to be complete. Some transitive verbs take sentential arguments as one of their essential arguments.

Verbs which take One Essential Arguments

There are verbs in English which need only one essential argument. These verbs are further subdivided according to the type of θ-roles their arguments take. Some verbs take only agent; some take theme/ patient or experiencer as their essential argument. Consider following examples:

In (18-21), the verb to run needs agent argument i.e. the boy, the verb to fell needs theme argument i.e. the tree, the verb to die needs patient arguments i.e., the boy and the verb to sneeze needs experience argument i.e., the child.

The Verbs which take Two Essential Arguments

There are verbs in English which need two essential arguments in their θ-grid to form a well- formed sentence. For example:

The Verbs which take Three Essential Arguments

There are verbs in English which take three essential arguments in their θ-grid to generate a well formed sentence. For instance:

The Verbs which take a Sentential Argument

There are verbs which take a sentential argument in their theta grid. Such verbs are further classified in two categories.

The verbs which need two argument i.e agent and theme.
The verbs which need three arguments i.e. agent, theme and experiencer.

Theme argument in both types of verbs is sentence or clause. For instance:

Theta Absorption (θ-Absorption)

An argument that is otherwise essential but is nevertheless omitted from the surface of the sentence and still the sentence is complete represents the case of theta absorption. In other words an essential argument is absorbed in the verb if it can be implied from it. Overt object complement is not needed on the surface of a sentence if it is absorbed in the verb. The verbs in which θ-absorption takes place are basically transitive. Consider following example:

31a. Raj is playing football.

31b. Raj is playing (a game).

In (31b), the theme argument i.e. a game has been absorbed in the verb play not a football. The arguments which are generic in nature or lacks specification can be absorbed in the verb such as something readable, something edible etc. The argument can be dropped from the surface if it can be understood from the context. The verbs in this regards vary language to language. In Hindi/Urdu we can drop the argument if it is contextually specified but in English we have to put ‘it’ at the place of the argument being dropped. For instance:

In (32b), the sentence has all the three arguments -agent ‘Raj’, theme ‘fruit’ phal, and source i.e. ‘fruit shop’ phal ki: duka:n. The construction of both English and Urdu/Hindi are correct. Sentences in (32c) have two arguments- agent ‘Raj’ and theme ‘fruit’ phal,. The source argument has been dropped in both the constructions of English and Hindi/Urdu and the sentences are correct.

In (32d) only agent argument i.e. Raj is present in the sentences of English and Hindi/Urdu both. English sentence of (32d) is not correct with agent ‘Raj’ only but its Hindi/Urdu equivalent is correct if we use such sentence in reference to earlier context. In English we will have to put ‘it’ at the place of the theme argument in order to make it correct i.e., Raj is buying it.

Chomsky (1981, 1986a) has coined the term θ-absorption for this phenomenon. It was Gruber (1976:11-12) who for the first time talked about this phenomenon in the name of Incorporation. The phenomenon of incorporation is not needed to discuss here so I am avoiding it. For example:

Summary

To sum up, the concept of the theta theory answered several question remained unanswered in earlier grammars. The notion of θ-criterion provides us a base to generate well formed sentences. The number of underlying theta role is the same in all languages but as languages differ in regard to their capacity of theta absorption, the surface manifestation of NPs with theta role may differ from language to language. Thus the theta theory/theta criterion accounts for all types of verbs and their possible arguments more adequately than the traditional division of verbs into intransitive and transitive does. Theta absorption states that an argument may not be in the surface structure of a sentence and still it is grammatical. It suggests that the omitted argument has been θ-absorbed. Such θ-absorbed argument can be recovered by analyzing the predicate itself.

you can view video on Predicate and Argument Structures

References

Baker, Mark C. 1988. Incorporation Chicago: The University of Chicago Press.
Chomsky, Noam. 1981 Lectures on Government and Binding. Dordrecht: Foris.
Chomsky, Noam. 1986a. Knowledge of Language: its Nature, Origin and Use. New York: Praeger.
Grimshaw, Jane. 1990. Argument Structure. Cambridge, MA:MIT Press.
Gruber, J.S. 1965(1976). Studies in Lexical Relations. (Doctoral Dissertation) Cambridge, Mass: MIT. Reprinted in 1976 as a Part of Lexical Structure in Syntax and Semantics. Amsterdam: North Holland.
Rizzi, L. 1990. Relativized Minimality. Cambridge. Mass: MIT Press.