1
PART ONE: INTRODUCTION
1. Rationale
Passive sentences are typical structures in Indo-European languages. A lot of
research about the passive phenomenon has been done, providing a wide variety of
features of passive structures cross-linguistically (e.g. Keenan, 1976; Freidin, 1975;
Permutter, 1978; Zaenen, 1983; Langacker, 1990; Tomlin, 1995.) Each school of
linguistics has displayed its own ways of analyzing the passive structures, thus
posing trouble for learners and for novice researchers in settling a standpoint in
grasping and understanding the fundamental semantic and syntactic structure of this
phenomenon in languages.
The following example illustrates this point. Sentences (a) and (b) below
describe a state in which a male person carries a box from one place to another:
(a) He moved the box into the room.
(b) The box was moved into the room (by him.)
Both sentences, (a) and (b), refer to one event with the same actor. Syntactically,
sentence (a) differs from sentence (b) in that the subject/doer in (b) is optional or
left out entirely, or optionally expressed obliquely via a by-phrase, and that the
object in (a) serves as the subject/patient in (b.)
Traditional grammar treats passive voice as the change of the morphology in
verbs, with the inversion of the subject and object of the sentences. Then, (b) can be
seen as a new version of (a) with a change in the verb „move‟ and an exchange of
location in the subject and object of the original sentence (a).
It has been argued in transformational grammar that passive sentences derive
from a D-structure form of so-called active sentences. It is the NP-MOVEMENT
(c.f. Radford, 1988:420) that brings to life passive sentences. Generative grammar 2
offers an analysis of the change from active to passive structures, attributing the
shift through a „multilevel syntagmatic structure‟ in which the „deep structure (or

study of language shifts from external to internal, mentalist aspects of the human
language. Researchers in cognitive science argue that the linguistic structure is a
direct reflex of cognition in the sense that a particular linguistic expression is
associated with particular ways of conceptualizing a given situation (Lee, 2001: 1).
They believe that there exists an interrelationship between thought and meaning. It
is from this point of view that passive sentences are seen from a different angle.
The root of the differences mentioned above originates from the theory of
meaning. Generative grammar claims that the structure of linguistic expressions is
determined by a formal rule system that is largely independent of meaning. Or
rather, syntax is independent from meaning while cognitive linguistics claims that
language is not an autonomous cognitive faculty (c.f. Croft, 2004.)
It is therefore absolutely necessary for researchers as well as language learners to
take a standpoint in analyzing languages. In particular, a thorough understanding
and fully developed arguments for the explanation of the structure are crucial.
A note should be taken here regarding the title of the dissertation. Since the word
„perspective‟ in cognitive semantics embraces more than one meaning and has sub-
notions hanging with it, this word used in the title may result in misunderstanding.
However, the word „perspective‟ in the title is used in purpose. The meaning of the
word „perspective‟ the author wishes to express is that of „view point,‟ „standpoint,‟
„angle,‟ or the like. With the title as it is, the author wishes to analyze the passive
structures from the cognitive semantics point of view. More than that, the author
wishes to bring a metaphor of the word „perspective‟ to the title, hoping to create a
profound and thoughtful meaning of the title of the dissertation. 4
2. Aims of the study
The study aims to provide a critical analysis of three major theoretical
approaches of explaining language phenomena. More particularly, different
approaches of linguistics, e.g. traditional grammar, transformational-generative

cognitive linguistics.
Regarding passive voice and structures to be analyzed, the study restricts itself to
structures with passive form in the main verb, i.e. the passive form of the utterance
is in the head verb, leaving behind structures consisting of passive forms in other
parts of the sentence, e.g. in the to infinitive, such as need to be repaired. Also, most
sentences used as illustrations come from everyday language, thus setting aside
extensive complicated sentences found in academic texts or elsewhere.
It should be noted here that it is not possible to apply one notion of cognitive
linguistics, such as Figure and Ground, in analyzing all kinds of passive structures.
This means that for each sentence, we have different ways of approaching and
applying proper notions for explaining. This is due to the fact that language
phenomena are various, and as a result, the way to deal with them varies
accordingly. Consequently, it is impossible to apply one idea to the explanation of
all features of language. However, the study is not going to categorize or classify
which kinds of sentences are applicable for which types of notions since it is not the
goal of the study. The last chapter in Part II is an attempt of assembling as many
notions as possible in analyzing the phenomenon of passive structures, thus
providing a broader view on this feature of language - passive structures. In short,
the study will focus on the facets that are applicable in analyzing passive structures,
not an overall presentation of the theory of cognitive semantics. 6
4. Research questions
Passive voice in English, or in Indo-European languages in general, has been
widely discussed from different points of view. Different papers present different
discussions on this language structure. This study attempts to seek answers to the
issues that have not been adequately discussed by many researchers from different
schools of linguistics. The questions that the study will investigate will be as
follows:

different specific situations in daily life. Sentences used as examples are cited from
either reliable sources or real life situations. All the examples from real life
situations, though not recorded and not put in the appendices, were sent to English
speaking people to test whether these sentences have a high acceptability in the
English language. More significantly, two native English speakers have been asked
to do the proof reading, providing high reliability and validity of the examples used
in the work.
5.2. The research design is presented as follows. First, in order to establish a
strong argument for the work to be done, all major points of views and the methods
of analyzing passive structures are critically illustrated. Then a framework is set up
from the light of cognitive semantics, i.e. a framework set up from four major
notions of cognitive semantics. Each of these four notions is presented with details
and illustrations, with which active and passive structures are analyzed and
explained. After that, an integration of these four notions is founded, from which
passive structures are analyzed from the overall examination of cognitive semantics. 8
This dissertation is divided into three parts. The first one is, as a rule of thumb,
the Introduction, which provides the rationale, aims, scope and research questions
of the study. This rationale provides arguments for the necessity of carrying out the
research; the aims presents the goals of the study and the scope restricts the study
into a certain aspects and leaves behind others for other future research.
The second part of the dissertation, the Development, presents a thorough review
of the literature about passive voice, in English and Vietnamese. More importantly,
a theoretical framework based on major notions of cognitive semantics is
established, providing foundation for the discussion and analysis of the passive
voice from the light of the cognitive semantics.
In part II, there are five chapters. Chapter 1 reviews major arguments and
analysis of passive voice from different schools of linguistics other than cognitive


10
PART TWO: DEVELOPMENT
In this Part, a review of theoretical foundation is done in the first half, serving as
a background for the study to be carried out in the rest of the part. This part consists
of five chapters. The first chapter deals with the arguments about passive voice
from three major linguistic theories, i.e. traditional grammar, transformational-
generative grammar, and functional grammar.
The first half of Chapter 2 displays the author‟s understanding of cognitive
linguistics in general and cognitive semantics in particular. This chapter confines
itself to the major points which will be applied to the analysis of passive voice in
the following chapters. The chapter starts with a definition of technical terms which
will be applied in the discussion; and it is followed by a general view on cognitive
linguistics. Then the chapter continues with an illustration of my understanding of
cognitive semantics, which serves as foundation for the study.
The rest of Chapter 2 presents an intensive review of major tenets of cognitive
semantics which have been applied in analyzing linguistics expressions. The chapter
displays four major notions of cognitive semantics, providing a theoretical
framework in the discussion of the phenomenon of passive voice in the following
chapters. The notions include Perspectives, Figure and Ground relation, Frame
Semantics and Encyclopedic Semantics, and Prototype.
Chapters 3, 4, and 5 are the core of the study. Chapter 3 applies the four major
notions of cognitive semantics, which have been displayed so far, into analyzing
passive structures. The discussion of passive structures are based on the major
tenets of cognitive semantics.
Chapter 4 deals with the passive structures in Vietnamese. Two major arguments
about whether or not there exist passive structures in Vietnamese are presented.
Then the chapter continues with the implementation of prototype theory in
Sentences in 1 refer to the same event, and according to traditional grammar,
they have the same meaning(s). The reasons of the passive usage (e.g. in 1.b or 1.c)
is attributed to very superficial reasons, e.g. because it is not necessary to mention 13
the doer, not knowing the doer, the doer being „people,‟ etc. (c.f. Thomson &
Martinet, 1960; Sinclair, 1990.)
The difference between the active sentence 1.a and the passive sentence 1.b or
1.c is superficially explained as a change in the morphology of the verb which must
be transitive; and the object of the sentence in active voice becomes the subject of
the passive counterpart.
The fact is that in many cases meaningful active sentences with transitive verbs
become abnormal if they are transformed into passive voice. For example, we can
say that „The shoes do not fit me‟, but rarely or never do we hear a sentence like „I
am not fitted by the shoes.‟ Moreover, it is undeniable that when one wishes to use
a different linguistic expression, she/he does want to convey a different meaning in
her/his utterance. Consider the following pair of sentences and see how the active
differs from the passive.
2. a. That man kicked the girl.
b. The girl was kicked by that man.
The sentences do not only differ syntactically but also semantically. While 2.a
may express the writer‟s intention to draw readers‟ or listeners‟ attention on how
brutal that man is, or rather on the vicious nature of the man, neither on the girl nor
on her injuries. On the other hand, with the sentence 2.b in passive form, the writer
obviously targets at his audience to sympathize with the girl and the way she has
been treated by the man. This example is a convincing illustration of the fact that
we would not have a profound understanding of the passive voice if we analyze and
interpret the meanings of passive and active sentences from the perspective of
traditional grammar. As a result, we need a different approach to deal with this

as the following:
3. The car will be put
__
in the garage.
(Radford, 1988:420) 15
And the NP [the car] is subsequently moved out of there into the italicized
preverbal subject NP position. The process can be described schematically in 4
below:
4. [
S
[
NP
e] will be [
V
put [
NP
the car] in the garage.]
NP MOVEMENT
Transformational grammar claims that the subject the car in the passive sentence
3 originates at D-structure in the „post-verbal Object NP position‟ in 4. In other
words, the passive sentence 3 derives from a D-structure, the active sentence in 4
(c.f. Radford, ibid.)
1.3.2. One of the arguments for the analysis of passive structures proposed by
transformational grammar is attributed to a class of expressions called idiom
chunks. It claims that since there is a class of English Noun Phrases which are
highly restricted in their distribution, which means that these NPs, in their idiomatic
use, generally occur only in junction with some specific verbs. The italicized NPs in

__
to the initial placed functioned subjects
in the sentence. This means that these NPs are restricted to occurring immediately
after specific verbs at D-Structure even though “they can move out of their
underlying position by movement rules like NP MOVEMENT” (Radford, 1988:
424.)
7. a. Little attention was PAID
__
to his report.
b. Little advantage was TAKEN
__
of the situation.
The NPs attention and advantage in 7.a and 7.b above are moved from their trace
marked
__
. Transformational grammar claims that these sentences originate from D-
structures which the NPs attention and advantage can only follow certain verbs, and
are called idiom chunk NPs.
1.3.3. Another argument posed by transformational grammar in describing the
passive voice is related to thematic relations which come from Chomsky‟s Theta
Theory. Theta Theory is a part of Government-Binding Theory which discusses the
assignment of theta roles for Arguments, i.e. NPs, in a sentence. In other words, 17
each Argument bears a particular thematic role (or theta-role or θ-role), e.g. Agent,
Patient, Experiencer, Instrument (c.f. Chomsky, 1981; Radford, 1988)
Regarding passive sentences, Theta theory states that the active objects, i.e.
objects of the active sentences, “play the same thematic role as the corresponding
passive subjects” (Radford, 1988: 424.) For instance, in the active sentence 8.a

NP I VP
e will AUX V
be V NP NP
sent Alice a new laptop.
NP
MOVEMENT

1.3.4. One more view from transformational grammar initiated by Wasow (1977)
is that passive voice is seen in two different types: adjectival and verbal. The
adjectival passives show the adjectival nature of passive participles which are
pronominal and complement to certain verbs, e.g. act, look. Some examples of the
adjectival passive are as follows:
11. a. The






edpa
broken
int
box sat on the table.
b. John







not as verbs, which leads him to the conclusion that passives can also be seen as
adjectives. Wasow argues that the rule that derives adjective passives must be a
lexical rule if it changes a verb into an adjective. The second type of passive,
according to Wasow, is called verbal passives which are derived by a
transformational rule. The rule Wasow mentions has just been analyzed previously.
Regarding the unrelatedness between transitivity and passivization, several
scholars state that it lies in the semantic properties of the transitive verb, i.e. the
affectedness of the Patient, the change the Patient undergoes by the force of the
Agent (c.f. Sudo, 2000.) In other words, when there is no change in the patient,
there is no possibility of passivizing an active sentence. Bolinger (1977) illustrates
the following examples:
14. a. John Smith deserted the army.
b. ? The army was deserted by John Smith.
15. a. All the generals deserted the army.
b. The army was deserted by all the generals.
According to Bolinger, the army, the Patient, is not affected if only Smith deserts
the army. However, if all the generals do, the army will definitely be affected, thus 20
causing a significant change in the structure of the army. Therefore, sentence 14.b
does not show a high acceptability, and it sounds unnatural to English speakers.
Bolinger‟s explanation seems to be logical to most of the cases on transitivity.
However, a number of different cases when there is not a high acceptability for the
passivization can not be attributed to the affectedness of the Patient. Take sentence
16 for example.
16. a. He has to park parallel the car for her.
b. ? The / Her car had to be parked parallel.
There is no question for the idea of affectedness to the car since at least when the
man parks parallel the car, he changes/affects the car in terms of space; but sentence

stratum. In other words, according to Perlmutter and Postal, “passive is a
„promotional‟ phenomenon, whereby a direct object nominal at one level is a
subject nominal at a later level” (Shibatani, 1985.)
Although Perlmutter and Postal have provided evidence for their arguments, they
have been criticized for the fact that not all passives show a promotion of a direct
object despite the fact that subject demotion occurs, e.g. impersonal passives.
Shibatani (ibid.) shows that the approach by Relational Grammar is “too restricted
to account for the patterns of distribution which a passive morphology exhibits.”
1.3.6. Major arguments from the perspectives of distinguished linguists of the
school of generative grammar have been presented. Briefly, generative grammars
have shown three basic features. One of which is that the synonymy is handled
through a multilevel syntagmatic structure for passive clauses, in which the deep
structure (D-structure) is basically active in organization. The second feature is that
the objects of by in passive structures are considered as a „demoted deep-structure
clausal subject‟ (Langacker, 1990:127.) The third primary feature is that “all or 22
some of the grammatical morphemes marking the passive construction (by, be, and
the perfect participle inflection) are meaningless entities, with purely formal or
syntactic function” (Langacker, 1990: 127.)
In other words, the generative approaches to the passive focus primarily on the
lexical-semantic and structural characteristics of the construction. These arguments
have triggered many debates for the last few decades. The drawbacks of the
approaches of generative grammar will be discussed later on in this study.
1.4. Passive voice from the perspective of functional grammar
1.4.1. Functional Grammar (henceforth FG) is an approach that analyzes the
grammatical categories in terms of their communicative functions. Language is seen
as a system of communication, and grammar is analyzed with an attempt to discover
how it is organized to allow speakers and writers to make and exchange meanings

attitudes.
The third type of meanings in FG is textual meaning. One sentence may have
different meanings in different contexts. In other words, language is used to
organize messages so that they fit with other messages around them; textual
meanings indicates the ways in which language is organized in relation to its
context (c.f. Halliday, ibid., Lock, ibid., Downing, 1992.)
1.4.2. Regarding passive voice, Halliday (ibid.) states that in clauses containing
Agents, i.e. doers, the clause is in passive voice if the Medium, i.e. the key figure
through which the process is actualized, and without which there would be no
process at all, serves as subject and the Agent as the adjunct. He claims that there
are two major reasons for choosing passive voice: First, to get the Medium as
Subject, and therefore as unmarked Theme, and second to make the Agent either (i)
late news, by putting it last, or (ii) implicit, by leaving it out (page: 168 – 169.)
Lock states that “the selection of passive voice allows the speaker or writer to
thematize participants,” (1996: 233) such as Goals, Recipients, without producing 24
marked Themes. Take for example the following pair of sentences, in sentence 17.b,
the Goal (the statue of David) is thematized, but as it is to the Subject, the Theme is
unmarked:
17. a. Michelangelo finished the statute of David in 1504.
b. The statute of David was finished (by Michelangelo) in 1504.
Another reason for the selection of passive voice, according to Lock (ibid.), is
motivated by “the desire not to switch the Theme,” or “to maintain continuity”
(Downing, 1992: 255;) for example:
18. She put her head out of the window and was struck full in the face by
a snowball.
(Lock, ibid.)
In 18, the Theme of the second clause is still She although it is ellipsed. The

e. The documents have been thrown away. {to highlight Predicator}
f. I‟m afraid the Fax hasn‟t been sent. {to be polite or to avoid
blaming someone}
Lock (ibid.) further presents some possibilities that determine the voice section.
One of which is to focus on a verb, in other words, to leave the end-focus to the
verb which is stated last, e.g. This must be seen to be believed; Moreover, passive is
also used in cases the Adjunct, or Complement, or the Prepositional Object in the
sentence receives the important end-focus, e.g. He was taken to jail; The letter has
been sent unstamped; or The retiring chairman was presented with a golden watch.