THE—282$

.‘\

lllllllllllllllllllllllllllllllllllllllllllllllllllll

3 1293 01770 067

 

Michigan State
, University

LIBRARY 1
I

 

This is to certify that the

thesis entitled

POLARITY SENSITIVE COPULAR VERBS IN THAI

“0

presented by '

Boonjeera”Chiravate
has been accepted towards fulfillment

of the requirements for

M.A. degree in Linguistics

ﬂ

Major professor

Date /3 QM/ M9

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

 

PLACE lN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

Jet.

""09 We;

 

 

 

      

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ma animus-Au

POLARITY SENSITIVE COPULAR VERBS IN THAI
By

Boonjeera Chiravate

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF ARTS
Department of Linguistics and Germanic, Slavic, Asian and Afn'can Languages

1 999

ABSTRACT
POLARITY SENSITIVE COPULAR VERBS IN THAI
By

Boonj eera Chiravate

Polarity sensitive items (PSIs) are expressions which can only occur in a positive
environment, in the case of positive polarity items (PPIS), or a negative environment, in
the case of negative polarity items (NPIs). In English the PPIs are expressions such as
scades (of money) and the NPIs are expressions such as (sleep) a wink. As there is a pair
of copular verbs in Thai—one PP] and one NPI, this thesis investigates the restricted
properties of these polarity sensitive copular verbs. Since these polarity sensitive copular
verbs in Thai do not carry semantic meaning of the type that the P815 in English do,
earlier studies which have been done on P815 have not resulted in an adequate
explanation. This thesis suggests that the explanation for the properties of polarity

sensitivity of the copular verbs in Thai might involve aspectual facts.

ACKNOWLEDGEMENTS

Profound appreciation in the preparation of this thesis goes to Professor Barbara
Abbott, my advisor. Her advice has inspired my fascination for linguistics and guided
my way through the whole thesis. I am grateful to the other members of my committee,
Professor Yen-Hwei Lin and Professor David Lockwood who gave me useful comments.

I have beneﬁted from discussion with SaisuneeVisonyanggoon who always gave
me valuable suggestions. I also would like to thank many friends in the department of
Linguistics for their generosity and support.

Finally my gratitude goes to my mother; her love and understanding gave me the

strength to ﬁnish this work.

iii

TABLE OF CONTENTS

CHAPTER 1
INTRODUCTION

CHAPTER 2

LITERATURE REVIEW
2.] A Summary of Research on PSI Licensing Problem
2.1 Israel (1996)
2.2 Summary

CHAPTER 3
POLARIT Y SENSITIVE COPULAR VERBS IN THAI
3.1 Polarity Sensitive Copular Verbs in Thai
3.2 The Problem of Polarity Sensitive Copular
Verbs in Thai
3.3 The aspectual Distribution of Polarity Sensitive
Copular Verbs in Thai
3.3.1 The Difference of Meaning between the Copular
Verbs in Thai
3.3.2 Restriction on Aspect: Comparing the Copular
Verb in English with the Copular Verbs in Thai
3.3.3 Restriction on Aspect: Comparing the Copular
Verbs with other Stative verbs in Thai
3.3.4 Summary of the Discussion on Aspect
3.4 An Alternative Approach to the Copular Verbs in Thai
3.4 Conclusion

CHAPTER 4
CONCLUSION

BIBLIOGRAPHY

iv

18
21

23
23
25
32
32
37
47
49
51
52
54

57

CHAPTER 1

INTRODUCTION

The restriction on the distribution of polarity-sensitive items (P815) is one of the
long standing problems in syntax and semantics. The expressions which can only occur
in a positive environment are positive-polarity items (PPI)l and the expressions which
can only occur in a negative environment are negative-polarity items (NPI). The class of
PPI in English includes some, already, pretty (AD V), etc., while the class of NPI includes
any, ever, yet, much, etc. Generally PPIs are unacceptable in negative sentences, as are
NPIs in positive sentences.

(1) a. *1 am not pretty happy with it.
(PPI in negative sentence)
b. I am pretty happy with it.
(PPI in positive sentence)
(2) a. *John has found his wallet yet.
(NPI in positive sentence)
b. John has not found his wallet yet.
(NPI in negative sentence)
In this thesis, I discuss a pair of copular verbs in Thai- one PPI (khu) and one NPI

(chat). The aim of this thesis is to investigate the restricted properties of these copular

 

1 The use of the term “positive” in this way is established in the literature but there is also
an argument for the use of the term “aﬁinnative” in contrast to “negative.”

verbs. The investigation will attempt to give an explanation for the distribution of the
polarity sensitive copular verbs in Thai.

There are three copular verbs in Thai: pen, khu and chat.2 As the data below
shows, pen is a polarity-insensitive, but khu and chat are polarity-sensitive. In particular,
khu is a PPI which has chat as a parallel NPI.

(3) a. John pen hmo
John be doctor
“John is a doctor.”
b. John mai pen hmo
John not be doctor
“John is not a doctor.”
(4) a. John khu hmo
John be doctor
“John is a doctor.”
b. *John mai khu hmo
John not be doctor
“John is not a doctor.”
(5) a. *John chat hmo
John be doctor

“John is a doctor.”

b. John mat chat hmo
John not be doctor
“John is not a doctor.”

Earlier studies which have been done on P813 have not resulted in an adequate
explanation. Most of them cannot account for the case of the polarity sensitive copular
verbs in Thai.

The theory of P815 should be able to answer two questions: the licensing question
and the sensitivity question. While the licensing question deals with what makes certain
contexts license polarity sensitivity, the sensitivity question involves what makes certain
forms sensitive to these contexts. Also logically the solutions for these two questions
should be related.

Much of the previous research has focused on the licensing problem. And since
the licensers for NPIs can be speciﬁed as environments containing negation while the
licensers for PPIs can be speciﬁed only negatively, in terms of absence of a negative
element, the research on P815 has focused on NPIs, in particular the NPI licensing
problem.

However, it seems that it is the sensitivity question, rather than the licensing
question that presents a bigger problem in Thai. Basically the NPI in Thai is licensed by
negation. However, comparing to English polarity sensitive expressions, such as budge
an inch, lift a ﬁnger, drink a drop, etc., Thai polarity sensitive expressions are copular

verbs. They do not cany semantic meaning of the type that the English PSIS do. So any

 

? I am using the transcription system of Royal Institute, 1982.

explanation which relates the semantic features that the P815 carry to the occurrence of
P815 does not work for the polarity-sensitive copular verbs in Thai.
What seems to be crucial is the fact that Thai polarity sensitive copular verbs
Show some restriction on aspect. For example, while the polarity-insensitive verb pen
allows the presence of the particles expressing tense, the polarity sensitive verbs khu and
chat do not.
(6) a. John cha pen hmo
John FUTURE be doctor
“John will be a doctor.”
b. John khoey pen hmo
John PAST be doctor
“John was a doctor.”
c. John cha mat pen hmo
John FUTURE not be doctor
“John will not be a doctor.”
(I. John mai khoey pen hmo
John not PAST be doctor
“John was not a doctor.”
(7) a. *John cha khu hmo
John FUTURE be doctor

“John will be a doctor.”

b. *John khoey khu hmo
John PAST be doctor
“John was a doctor.”
(8) a. *John cha mai chai hmo
John FUTURE not be doctor
“John will not be a doctor.”
b. *John mat khoey chai hmo
John not PAST be doctor
“John was not a doctor.”
In this thesis I suggest that the explanation for the distribution of the polarity sensitive
copular verbs in Thai involves these aspectual facts.
The reminder of this thesis consists of 3 chapters. Chapter 2 summarizes research
related to polarity sensitivity and aspectuality. Chapter 3, examines the distribution of

the copular verbs in Thai and attempts to account for the distribution of those which are

polarity sensitive. Chapter 4 states the conclusions of the thesis.

CHAPTER 2

LITERATURE REVIEW

There do not seem to be earlier studies on the Thai language that focus on the
polarity sensitivity of the copular verbs. Grammars of Thai, such as Uppakitsinlapasam
(1964) do not provide any explanation about them.

This chapter provides a review of the research relevant to polarity sensitivity.
Most of the research concerns polarity sensitive items in English. Section 2.1 presents a
summary of the research done by Ladusaw (1980, 1982, 1983), Van der Wouden (1994),
and Linebarger (1981, 1987, 1989) which focuses on the licensing problem, in particular,
what makes certain contexts license polarity sensitivity.3 Section 2.2 summarizes the
research done by Israel (1996) which focuses on the sensitivity problem, in particular,

what makes certain forms sensitive to those contexts. Section 2.3 concludes the chapter.

2.1 A Summary of Research on PSI Licensing Problem

Since the earliest work on NPIs (Klima 1964), it has been recognized that

expressions which can license NPIs can be both overt negation like not, nobody and

never as shown in (9) and covert negation like few and rarely as shown in (10). In

 

3 Another research focusing on the licensing problem is the research done by Progovac
(1992, 1994). Her account is based on Binding Theory and accounts for the facts
concerning non-negative polarity licensing, the presence of NPIs in wh-questions and the
NPI licensing by the determiner only. Since the copular verbs in Thai do not occur in
these contexts, her account is not relevant to the copular verbs in Thai. So I do not
provide the summary of her account here.

addition, some environments such as an antecedent of conditionals and yes/no questions
allow an occurrence of NPIS as well, as shown in (1 l).
(9) a. I didn’t drink a drop at the party.

b. Nobody has found a helpful book yet.

c. They never contribute a red cent.
(10) a. Few students ever take the train to school.
b. Bill rarely budges an inch on such matters.
(11) a. Ifanyone has questions, please ask me.

b. Do you have any cats?

Ladusaw (1980, 1982, 1983) proposed that the expressions that can license NPIS
can be determined by semantic criteria. They are the class of expressions whose
meanings are downward entailing.4

Ladusaw’s analysis involves operators and arguments. A negation NOT is
considered to be an operator which takes a sentence as its argument. By taking NOT as a
sentence operator, a negated sentence like (12a) can be presented as (12b).

(13) a. John didn’t eat fruit.
b. not (John ate fruit)
OPERATOR ARGUMENT

By taking NOT as a sentence operator he can explain the ambiguity of sentences

of the following kind.

(13) John didn’t move because he was frightened.

 

4 Downward entailing is also known as monotone decreasing.

In (13), both the reading “It is because he was frightened that he didn’t move” and the
reading “It is not because he was frightened that he moved” are possible. According to
Ladusaw, this ambiguity arises from the fact that the negation NOT is an operator which
can take scope over its arguments in two different ways, as shown below.

(14) a. because (he was frightened) (NOT (he moved))

b. NOT (because (he was frightened) (he moved))

Given that the because clause is an adjunct clause and the clause modiﬁed by the
because clause is a matrix clause, in (14a) the negation NOT takes a narrow scope, to be
more speciﬁc, it takes scope only over the matrix clause but in (14b) the negation NOT
takes a wide scope, to be more speciﬁc, it takes scope both over the adjunct clause and
the matrix clause. So (14a) provides the ﬁrst reading “It is because he was frightened
that he didn’t move” and (14b) provides the second reading “It is not because he was
frightened that he moved.”

Apparently the interpretation of a sentence is the mapping between two sets, in
particular, the set consisting of sentences and the sets consisting of truth values. As for
each sentence we have assigned a truth value, the assignment is called a function. So an
operator is considered to be a function.

As Ladusaw claims that the NPI licensers are expression whose meanings are
downward entailing, the property of downward entailingness can be expressed by the
following.

(15) A ﬁmctor f is downward entailing, iff if f (X), and Y is a subset of X, then f (Y)

So downward entailment is an entailment from a superset (a less restrictive term )

to a subset (a more restrictive term).

The negation NOT is regarded as a sentence operator which is downward
entailing with respect to its argument. He claims that NPIS are licensed in downward
entailing environments. Consider, ﬁrst, the following two propositions P and Q:

(16) a. John ate an apple. (P)

b. John ate fruit. (Q)

P entails Q. In particular, the direction of entailment is from the subset (the possible
world expressed by “John ate an apple”) to the superset (the possible world expressed by
“John ate fruit”). This is not downward entailment. Consider, on the other hand,
sentences with NOT as a sentence operator, in particular, the negated version of ( 1 6a )
and (16b)

(17) a. John didn’t eat fruit. (NOT O)

b. John didn’t eat an apple. (NOT P)

Here the direction of entailment is reversed. NOT Q entails NOT P. The entailment is
from the superset (the set of possible worlds expressed by “John didn’t eat hit”) to the
subset (the set of possible worlds expressed by “John didn’t eat an apple”). So this is
downward entailment. Since downward entailment environment licenses NPIS, the NP]
anything will be licensed, as in the following.

(18) John didn’t eat anything.

Therefore, if an expression licenses inferences from superset to subset, it is downward
entailing. And it can license the occurrence of an NPI.

Ladusaw claims that a negation NOT is only a subclass of an operator whose
meaning is downward entailing with respect to its argument. A determiner such as no is

also viewed as an operator whose meaning is downward entailing with respect to its

argument and which can license an NPI is its argument. While the negation NOT takes a
sentence as its argument, the determiner no takes a common noun phrase (CNP) as its
argument, as Shown by the following.
(19) a. No man

b. No (man)

OPERATOR ARGUMENT
According to him the truth value of No man is the result of the fact that the CNP man has
to be interpreted with respect to the operator no. The determiner no is downward
entailing expression, as P entails Q in the following example.
(20) No man walks (P)
No father walks (Q)

So the determiner No can license an NPI in its argument. The NPI ever is licensed in the
following examples.
(21) No person who has ever visited New York has returned to it.

The proposal of Ladusaw (1980, 1982, 1983), however, has some problems.
Consider, for example, the following;
(22) a. At most three students like sports. (P)

b. At most three students like tennis. (Q)
The quantiﬁed NP at most three students is viewed as an operator. In (21) as P entails Q,
at most three students should be a downward entailing operator that can license an NPI in
its argument. However, while the NPI any is licensed, the NPIS yet and a bit are not

licensed as shown in (23).

10

(23) a. At most three students have played any Sports.
b. *At most three students have played sports yet.
c. *At most three students were a bit willing to play sports.

Van der Wouden (1994) accounts for this problem by categorizing the classes of
NPI licensers into three classes: monotone decreasing (MD) operators, anti-additive
operators and antimorphic operators. He gives the deﬁnition for each of the class as the
following.

(24) A function f is monotone decreasing (downward entailing) iff f (X or Y) -> f (X)

and f (Y)

[= van der Wouden 1994, ex. 9]

(25) A function f is anti-additive iff f (X or Y) <-> f (X) and f (Y)

[= van der Wouden 1994, ex. 13]

(26) A functor f is antimorphic iff f (X) and f (Y) <-> f (X or Y) and f (X) or f (Y) <->

f (X and Y)

[= van der Wouden 1994, ex. 16]

Few children is a monotone decreasing noun phrase and no children is an anti-
additive operator and negation not is an antimorphic operartor, as shown by (26) and (27)
and (28) respectively.

(27) Few children sing or dance -> few children sing and few children dance.

[= van der Wouden 1994, ex. 10c]

(2 8) No children sing or dance <-> no children sing and no children dance

[= van der Wouden 1994, ex. 14a]

11

(29) a. Not sing and not dance <-> not (sing or dance) and

b. Not sing or not dance <-> not (sing and dance)

[= van der Wouden 1994, ex. 17a -l7b]

(29) can be illustrated as (30).
(30) a. Children do not sing and do not dance <-> Children do not sing or dance

b. Children do not sing or do not dance <-> Children do not sing and dance

He claims that different NPIS are associated with different licensers. Some NPIS
are so strong that they require only the strongest class of licensers, antimorphic operators.
Some NPIS are relatively less strong, so they can be licensed by both antimorphic
operators and anti-additive operators. And some NPIS are weak as they can be licensed
by antimorphic operators, anti-additive operators and the weakest class, monotone-
decreasing operators. Subsequently he classiﬁes NPIS in to three types; strong NPIS,
medium NPIS, and weak NPIS. To illustrate, consider the NPIS a bit, yet and any as the
representatives of three types of NPIS in the following examples (1994: p. 19)

Anti-morphic licensers:
(31) a. Chomsky wasn’t a bit happy about these facts

b. Chomsky didn’t talk about these facts yet

c. Chomsky didn’t talk about any of these facts

Anti-additive licensers:
(32) a. *No one was a bit happy about these facts

b. No one has talked about these facts yet

c. No one has talked about any of these facts

12

Monotone decreasing licensers:
(33) a. ‘At most three linguists were a bit happy about these facts

b. *At most three linguists have talked about these facts yet

c. At most three linguists have talked about any of these facts

The NPI a bit can be licensed only by anti-morphic license. The NPI yet can be
licensed by both anti-morphic and anti-additive licenses. The NPI any can be licensed
by, anti-morhic licenses, anti-additive and monotone decreasing licenses. His
assumption is that a bit is a strong NPI and yet is a medium NPI and any is a weak NPI.

So both Ladusaw (1980, 1982, 1983) and Van der Wouden (1994) capture the
phenomenon of NPIS by semantics. By these approaches, the NPI licensing depends on
the semantic properties of an expression and does not depend on context.

By considering the NPI licensing to be context-independent, there remains a
problematic case. Consider the following sentence.
(34) I was surprised that she contributed a red cent.

[ = Progovac 1994, ex. 39]
The NPI a red cent is licensed in this sentence. Assuming that be surprised that is a
licenser, according to Ladusaw (1980, 1982, 1983)’S account, be surprised that should be
downward entailing. However, it is not clear that be surprised that is downward
entailing because it is not clear that (35a ) entails (35b).
(35) a. John was surprised that Bill sang.

b. John was surprised that Bill sang a happy song.

13

Sentences like ( 34) are also problematic for Van der Wouden (l994)’s account. He notes
that his deﬁnition of monotone decreasing does not work completely in the case of
affective verbs. To illustrate, the following example has been given ( 1994: p.21);
(36) a. John regrets that anyone was injured

b. John regrets that Mary or Susan was injured

c. John regrets that Mary was injured and John regrets that Susan was

injured.

Although to regret licenses the NPI anyone in (36a), (36b) does not entail (36c). The
same thing seems to go for be surprised that. Although it licenses the NPI a red cent, it
is not clear that it is a monotone decreasing as it is not obvious that (37a) entails (37b).
(37) a. John was surprised that Bill sings or dances.

b. John was surprised that Bill Sings and John was surprised that Bill dances

Linebarger (1980, 1987, 1991) tries to account for the kind of sentences like ( 33).
She explains the distribution of the NPIS by syntax and pragmatics. She claims that an
NPI licenser must be a negation expression. A negation expression is either in a negative
sentence or in an implicature of a non-negative sentence. Her analysis consists of two
parts; Part A and Part B.

Part A deals with the case that an NPI is licensed by an overt negation. To
illustrate, consider the following sentences.
( 38) He didn’t move because he was frightened.
(39) He didn’t budge an inch because he was frightened.
While (38) is ambiguous, (39) is not. In particular, in (38) both the reading of “It is

because he was ﬁightened that he didn’t move” and the reading of “It is not because he

14

was frightened that he moved” are possible. However, in (39), only the ﬁrst reading is
possible.

To account for the absence in an ambiguity of the sentence containing an NPI like
(39), she proposed the Immediate Scope Constraint (ISC), which says that an NPI is
acceptable in a sentence S if in the LF of the S the representation of NPI is in the
immediate scope of the negation operator.

As (39) is a grammatical sentence, in the LP of (39), the NPI budge an inch must
be in the immediate scape of the negation not. She explains that the two LF candidates
for (39) will be (40a) and (40b), where S-l is the adjunct clause (the because clause) and
S-2 is the matrix clause (the clause modiﬁed by the because clause).

(40) a. NOT CAUSE (S-l , 8-2)
It is not because he was frightened that he budged an inch.

b. CAUSE (S-l, NOT S-2)

It is because he was frightened that he didn’t budge an inch.

In (40a) there is an intervention of the predicate CAUSE between NOT and 8-2.
So the NPI budge an inch in 8-2 is not in the immediate scope of the negative operator
NOT. So the reading that NOT takes scope over S-1 and at the same time licenses the
NPI in 8-2 is not possible. That is why (39) cannot be interpreted as "It is not because he
was frightened that he budged an inch.” But in (40b) the NPI budge an inch in 8-2 is in
the immediate scope of the negative operator NOT. And S-l is not in the immediate
scope of NOT at all. So that is why (39) has to be interpreted only as "It is because he

was frightened that he didn‘t budge an inch."

15

Therefore, Part A of her analysis says that an NPI will be acceptable in a sentence
S, if in the LF of the S the representation of an NPI is in the immediate scope of the
negation operator.

Part B seems to take care what is missed by Ladusaw (1979)’s account. It covers
the case of NPI in non-negative polarity contexts. The idea is that if there is some
negative implicature which is part of what the speaker is attempting to convey by saying
S, and in the LP of the implicated sentence the lexical representation of an NPI occurs in
the immediate scope of negation, then the S can license the NPI.

The following example from Linebarger (1981) is taken to illustrate this point:
(41) a. I was surprised that she contributed a red cent.

[= Progovac 1994, ex. 39]
b. IMPLICATURE
I had expected her not to contribute a red cent.
[= Progovac 1994, ex. 40]
In uttering (41a), the speaker is making an allusion to (41b) in which that NPI a red cent
is in the immediate scope of the negation not. So by virtue of implicating (41b), (41a) is
able to license an NPI.

Linebarger emphasizes that the negative implicature must be part of what the
speaker is using the sentence to convey. To illustrate the point, the following examples
have been given.

(42) a. Cows ﬂy more than he liﬁs a ﬁnger to help
b. If it’s rare for cows to fly, then he almost never lifts a ﬁnger to help.

[= Linebarger 1980, ex.42]

l6

(43) a. ‘The sun rises more often than he liﬁs a ﬁnger to help.
b. It it’s rare for the sun to rise, then he almost never lifts a ﬁnger to help.
[=Linebarger 1980, ex 43]

In both the LP of implicatures (42b) and (43b) the NPI lift aﬁnger is in the
immediate scope of the negation not, however, while (42a) is acceptable, (43a) is not.
Linebarger explains that this is because the availability of the implicatures is different.
The licensing implicature (42b) is informative enough that we might expect (42a) to be
used to convey it. In particular, since cows never ﬂy, it tells us that he will never help.
So the sentence will be true only if he never helps. But the licensing implicature (43b) is
so uninformative that it is unlikely that (43a) would be used to convey it. Since the sun
rises everyday, it tells us that he will be helping any amount that is not greater than
everyday. So the sentence will be true in cases such as that he helps every week, that he
helps every month, etc.

In the LP of the implicature (41b) the NPI a red cent is in the immediate scope of
the negation not and the implicature (41b) is informative enough that we might expect
the speaker to convey it by uttering (41a). So sentence (41a) is an acceptable sentence.

In summary, the two leading approaches to negative polarity are Ladusaw( 1980,
1982, 1983), and Linebarger (1980, 1987, 1991). Ladusaw’s account, which is based on
semantic entailment, seems to have a number of problems. The account proposed by
Van der Wouden (1994), which is in favor of semantic approach, seems to ﬁx some of
the problems, but still cannot account for the case of NPI in the absence of downward

entailing (monotone decreasing) expression.

17

Linebarger (1980, 1987, 199l)’s account seems to be more problem-free. By
proposing that the NPI licensing applies at the LF level and at the implicature, she can
account for all the problematic cases.

However, these accounts of negative polarity, while answering the licensing
question, in particular, what makes certain contexts license polarity sensitivity, do not
provide an answer for the sensitivity question, in particular, what makes certain forms
sensitive to these contexts. In the next section I summarize another account of polarity

sensitivity which can answer both the licensing and the sensitivity question.

2.2 Israel (1996)

Israel (1996) proposed a lexical semantic explanation for the distribution of
polarity sensitive items. Though focusing on the sensitivity question, his account
answers the licensing question as well. He considered the problem of polarity sensitivity
as lexical semantics. He claims that PSIS are speciﬁed for two semantic features:
quantitative value and informative value. The interaction of these two features makes
them sensitive to certain contexts.

His idea is that some words range in terms of strength, for example, the evaluative
terms excellent, good and 0.1:. range in the degree of perfection. The word excellent
expresses the higher degree of perfection than the word good and the word good
expresses the higher degree of perfection than the word 0.1:. So on the scale of perfection

the word excellent is encoding a higher quantitative value than either good or air.

18

Israel claims that PSIs encode quantitative value (q-value), in particular PSIs
encode either a high q-value or a low q-value. For example, on the scale of the amount
of sleeping a wink is designating a low q-value while much is designating a high q-value.

(44) a. Margo didn’t sleep a wink before her big test.
[=Israel 1996, ex. 10a]
b. Margo didn’t sleep much before her big test.
[=Israel 1996, ex. 10b]
On the scale of the amount of money scads is encoding a high q-value and a little bit is
encoding a low q-value.
(45) a. Belinda won scads of money at the Blackjack tables.
[=Israel 1996, ex.11a]
b. Belinda won a little bit of money at the Blackjack tables.
[=Israel 1996, ex.11b]

Israel claims that a sentence containing a PS1 implicitly refers to a norm. A norm
can be understood as a normal expectation in context. A sentence which is more
informative than a norm is an emphatic sentence and a sentence which is less informative
than a norm is an understatement sentence. So a norm will be entailed by an emphatic
sentence and an understatement sentence will be entailed by a norm.

When a particular word is conventionally associated with emphatic or
understating contexts, the word can be stereotyped as conveying an emphatic or
understating force. PSIs are the words of this kind. They are speciﬁed for informative
value (i-value), in particular, they are encoding either an emphatic force or an

understating force.

19

Israel assumes that an NPI which designates a low q-value is equipped with an
emphatic force when it is negated so it will produce an emphatic sentence, such as (44a).
An NPI which designates a high q-value is equipped with an understating force when it is
negated so it will produce an understatement sentence such as (44b). And it is backward
when polarity is reversed. A PPI which designates a high q-value is equipped with an
emphatic force so it will produce an emphatic sentence such as (45a). A PPI which
designates a low q-value is equipped with an understating force so it will produce an
understatement sentence such as (45b).

An NPI will produce an emphatic or an understating sentence only in the

environment that a lower position entails a higher position. To illustrate, consider the

following ﬁgure.

(46) --- The largest amount of sleeping
--- X2
--- X1

---- The smallest amount of sleeping
Both (44a) and (44b) will entail “Margo didn’t sleep the largest amount of
sleeping.” That is, NPIS will be licensed only in the environments that reverse the
direction of entailment in the model. That environment is apparently a negative
environment. That is why these forms are NPIS.
In contrast, a PPI will produce an emphatic or an understating sentence only in
the environment that a higher position entails a lower position. To illustrate, consider the

following ﬁgure.

20

(47) --- The largest amount of money
--- X2
--- X1
--- The smallest amount of money
Both (45a) and (45b) will entail "Belinda won the smallest amount of money."
That is PPIs will be licensed only in the environments that preserve the direction of
entailment in the model. And that environment is apparently a positive environment.
That is why they are PPIS.
So by assuming that PSIs are Speciﬁed for q-value and i-value which relates to the
notion of emphasis and understatement, and by deﬁning the notions of emphasis and
understatement in terms of entailments, both the sensitivity and the licensing problem

can be solved.

2.3 Summary

There seem to be two main questions that the theory of polarity sensitivity should
be able to answer: the licensing question (what makes contexts license polarity
sensitivity) and the sensitivity question (what makes certain forms sensitive to these
contexts). Also logically the solution for these two questions should be related.

However, apparently much of the research has focused on the licensing question,
in particular NPI licensing. Different approaches to NPI licensing have been proposed.
The two main approaches are Ladusaw (1980, 1982, 1983) and Linebarger

(1981 ,1987,1989 ). Ladusaw argues that the licensing of NPIS can be determined by the

21

semantic property of being downward entailing. According to Ladusaw, negations are
only a subclass of downward entailing expressions. Linebarger, on the other hand, argues
the distribution of NPIS has to be explained by the interplay between syntax and
pragmatics. For her, what licenses an NPI in a sentence S is a negaan which can be
either in the LP of the sentence S or in the LP of the implicature of the sentence S.

Israel (1996), on the other hand, offers a neat explanation based on lexical
semantics which can answer both licensing and sensitivity questions. His answer to the
one is related to the other. He claims that PSIs encode two semantic features,
quantitative value and informative value and the interaction of these two features causes
the polarity sensitivity.

In summary, different approaches have been proposed to account for the
phenomenon of polarity sensitivity. Each of the approaches has its strengths and
weaknesses. What all of them share in common seems to be an attempt to make the

generalization that can account for the mystery of polarity sensitivity in languages.

22

CHAPTER 3

POLARITY SENSITIVE COPULAR VERBS IN THAI

In this chapter, I investigate the restriction on the distribution of the copular verbs
in Thai and attempt to give an explanation for the polarity sensitivity of these verbs.

This chapter consists of 4 sections. Section 3.1 shows that there is a pair of
copular verbs in Thai-one PPI and one NPI. Section 3.2 shows how the data from the
copular verbs in Thai poses problems with respect to polarity sensitivity items. Section
3.3 examines the aspectual distribution of the copular verbs and discusses the possible
connection between the aspectual facts and the polarity sensitivity of the copular verbs.
Section 3.4 provides an alternative approach to the copular verbs in Thai. Section 3.5

concludes the chapter.

3.] Polarity Sensitive Copular Verbs in Thai

Copular verbs are verbs that make no semantic contribution to a sentence. The
verb be is considered a copular verb in English. Thai has three copular verbs: pen, khu
and chai. The distribution of the copular verbs involves polarity sensitivity. While the
copular verb pen can occur in both positive and negative environments, the copular verb
khu can only occur in positive environments and the copular verb chai can only occur in

negative environments.

23

 

(48)

(49)

(50)

(51)

a. Somehay pen prathan samakhom nakrian Thai
Somehay be president association student Thai
“Somehay is the president of Thai student association.”

b. Somehay mat pen prathan samakhom nakrian Thai
Somehay not be president association student Thai
“Somehay is not the president of Thai student association.”

a. Somehay khu prathan samakhom nakrian Thai
Somehay be president association student Thai
“Somehay is the president of Thai student association.”

b. *Somchay mai khu prathan samakhom nakrian Thai
Somehay not be president association student Thai
“Somehay is not the president of Thai student association.”

3. *Somchay chai prathan samakhom nakrian Thai
Somehay be president association student Thai
“Somehay is the president of Thai student association.”

b. Somchay mai chai prathan samakhom nakrian Thai
Somehay not be president association student Thai
“Somehay is not the president of Thai student association.”

Let us summarize the above observation as follows:

Pen is polarity-insensitive, but khu and chai are polarity sensitive, in particular

khu is a PPI which has chai as a parallel NPI.

24

3.2 The Problems of Polarity Sensitive Copular Verbs in Thai

Despite the fact that most of the research has focused on the licensing problem, in
particular what makes certain contexts license polarity sensitivity, the data from the
polarity sensitive copular verbs in Thai seems to pose more problems regarding
sensitivity, in particular, what makes certain forms sensitive to these contexts.

To begin with, the analysis by Ladusaw (1980, 1982, 1983) involves a sentence
operator and an argument. He claims that a downward entailing operator can license an
NPI in its argument. A negation is a subclass of the downward entailing operators and so
it can license an NPI.

Generally the NPI chai in Thai is licensed by the negation mai “not.” (52a) can be
presented as (52b) and (53a) entails (53b),so mai “not” is a sentence operator which is
downward entailing and can license the NPI chai.

(52) a. John mai chai hmo

John not be doctor

“John is not a doctor.”
b. mat (John chai hmo)
not John be doctor

OPERATOR ARGUMENT
(53) a. John mai chai hmo
John not be doctor

“John is not a doctor.”

25

b. John moi chai hmo ti chaidi
John not be doctor that nice
“John is not a nice doctor.”
So Ladusaw’s proposal seems to apply here.

In Thai there seem to be two words siathihnai and siamuarai which when put at
the end of a sentence can change the polarity of the sentence, as shown by the following
examples.

(54) a. John chai hmo siathihnai
John be doctor NEG
“John is not a doctor.”
b. John chai hmo siamuarai
John be doctor NEG
“ John is not a doctor.”
(55) a. John mat chai hmo siathihnai
John not be doctor NEG
“John is a doctor.”
b. John mai chai hmo siamuarai
John not be doctor NEG
“John is a doctor.”

So it seems that these words when put at the end of the clause, negate the whole

clause. In (54), they negate the clause John chai hmo “John is a doctor” and in (55) they

negate the clause John rnai chai hmo “John is not a doctor.”

26

In (55), as double negation produce positive meaning, the sentences mean “John
is a doctor.” The sentences in (55),therefore, can be represented as the following.
(56) a. siathihnai (mai (John chai hmo) )
NEG not John be doctor
OPERATOR OPERATOR ARGUMENT
b. siamaurai (moi (John chai hmo) )
NEG not John be doctor
OPERATOR OPERATOR ARGUMENT
So despite of the positive meaning of the sentences like (55), the account proposes by
Ladusaw still applies. The NPI chat is licensed by the downward entailing sentence
operator which is, in the case of the sentences in (55), the negation mai “not,” and the
other round of negation which turns the sentences into positive sentences does not effect
the NPI licensing.
Van der Wouden (1994) proposes that NPIS can be classiﬁed into 3 classes; weak
NPIS, meduim NPIS, and strong NPIS, according to their licensing requirement. The NPI
chai requires the strongest class of licensers, in particular the anti-morphic licenser, such
as (57). Neither an anti-additive licenser, such as (58) nor a monotone decreasing
licenser, such as (59) is sufﬁcient to license it.
(57) John mai chai hmo
John not be doctor

“John is not a doctor.”

27

(58) I“Mai mi khrai chai hmo
Not have who be doctor
“No one is a doctor.”

(59) *Yang mak hmo sam khan chai khon Thai
At most doctor three people be people Thai
“At most three doctors are Thai people.”

So the NPI chai should be classiﬁed as a strong NPI.

So far the accounts in favor of a semantic approach, in particular Ladusaw (1980,

1982, 1983) and Van der Wouden (l994),therefore, seems to be working for the NPI

licensing in Thai. The account based on syntactic conﬁguration proposed by Linebarger

(1980, 1987, 1981) also seems to apply.

According to Linebarger (1980, 1987, 1981), the NPI licenser must be a negation.

In particular she claims in the Part A of her analysis that for an NPI to be acceptable in a

sentence S, it has to be in the immediate scope of the negation in the LP of the sentence

S. This seems to apply for the case of the NPI copular verb in Thai. Consider the

following sentences.

(60) John mai dai pen hmo phro tuk phomae bangkhap (tae phro khao
John not become be doctor because PASSIVE parents force (but because he
yak pen eng)
want be self)

“John was not a doctor because he was forced by his parents (but because he

himself wanted to be a doctor. )”

28

(61) *John mat dai chai hmo phro tuk phomae bangkhap (tae phro khao
John not become be doctor because PASSIVE parents force (but because he
yak pen eng)
want be self)

“John was not a doctor because he was forced by his parents (but because he

himself wanted to be a doctor. )”

The LP representation of both (60) and (61) can be represented as (62), where 8-]
is the adjunct clause (the because clause) and 8-2 is the matrix clause ( the clause
modiﬁed by the because clause).

(62) NOT CAUSE (S-l, 8-2)

The forms pen and chai in bold face in (60) and (61) are located in 8-2 of (62) so
they are not in the immediate scope of the negation. As the form pen is not an NPI so the
sentence is still acceptable but the form chai is an NPI so the sentence is not acceptable.

According to Linebarger, for the form chai to be acceptable, it has to be, at the
LP, in the immediate scope of the negation, such as the following representation.

(63) CAUSE (S-l, NOT 8-2)

A sentence corresponding to this LF representation will be the sentence such as
the following.

(64) John mat chai hmo phro khao rian ek compiwtoe
John not be doctor because he study major computer
“John is not a doctor because he was majoring in computer science.”

In (64) the form chai which is in 8-2 is in the immediate scope of the negation mai “not”,

so, as predicted by Linebergar’s account, the sentence is acceptable.

29

In conclusion, the NPI chai in Thai is generally licensed by overt negation. Since
licensing by overt negation is the simplest case of NPI licensing, any accounts on NPI
licensing seem to apply.5

Rather than for licensing, Thai polarity sensitive copular verbs pose problems for
sensitivity. Israel (1996) offers an account focusing on the sensitivity problem, in
particular, what makes certain forms sensitive to those contexts. His account eventually
can answer both licensing and sensitivity problems. He claims that PSIs are speciﬁed
for two semantic features: quantitative value and informative value. The interaction of
these two features makes them sensitive to certain contexts. His explanation, however,

does not apply for the case of the polarity sensitive copular verbs in Thai.

 

5 There is, however, a case of NPI licensing in Thai, which seems to call for more
explanation, as shown by the following.
(i) chai wa John chop Mary
be that John likes Mary
“It is not that John likes Mary.”
(ii) Mai chai wa John chop Mary
Not be that John likes Mary
“It is not that John likes Mary.”
AS (i) means the same thing as (ii), the word chai “is” means mai chai “is not.” The

explanation for this is beyond the scope of this paper.

30

Copular verbs do not carry semantic meaning of the type that the English PSIs do.
Obviously they cannot be speciﬁed for q-value. For the i-value, Israel has shown that if
an item is speciﬁed for i-value, it will be limited to certain contexts, in particular, the
context that supports its i-value. As the distribution of the copular verbs in Thai shows
that they can only occur in certain contexts, it seems that the i-value must be involved.

However, Israel has shown that if an item can only occur in negative context, it is
its low q-value that will push it to produce emphatic sentence and it is its high q-value
that will push it to produce an understatement sentence. This is not what happened in the
case of PSIs in Thai. The item that can only occur in a negative context is not equipped
with any q-value, so it does not produce either emphatic or understatement sentences.
The same goes for an item that can only occur in a positive context. He has shown that if
an item can only occur in a positive context, it is its low q-value that will push it to
produce understatement sentence, and it is its high q-value that will push it to produce an
emphatic sentence. In the case of the PS1 in Thai, the item can only occur in a positive
context but it is not equipped with any q-value so it produces neither an understatement
nor an emphatic sentence.

Related to the explanation conceming the features that PSIs are speciﬁed for,
Israel answers the licensing question. He explains that the environments that allow items
like a wink and much to encode their features are actually the environments that a lower
position entails a higher position, which is a negative environment. The environments
that allow items like scads and a little bit to encode their features are actually
environments where a higher position entails a lower position, which is a positive

environment.

31

However, this is not what happened in the case of polarity sensitive copular verbs
in Thai. Since they are not equipped with those features. They do not locate on a scale.
So the environments that license them has nothing to do with the position and the
entailment.

In summary, the data from the polarity sensitive copular verbs in Thai does not
present a problem for licensing. The accounts on NPI licensing that have been proposed
apply for the case of NPI licensing in Thai. The data from the polarity sensitive copular
verbs in Thai, however, presents a big problem to the sensitivity. The previous account

of sensitivity fails to account for the case of polarity sensitive copular verbs in this

language.

3.3 The Aspectual Distribution of Polarity Sensitive Cepular Verbs in Thai

In this section I Show that the distribution of copular verbs in Thai involves aspectual
restrictions. This section consists of 4 parts. In 3.3.1 I summarize the research done by
Kuno and Wongkhomthong (1981) concerning the difference in meaning between the
copular verbs in Thai. Sections 3.3.2 and 3.3.3 examine the restrictions on aspect of the

copular verbs. Section 3.3.4 summarizes the discussion on aspect.
3.3.1 The difference of meaning between the copular verbs in Thai

It is helpful to begin an aspectual discussion of the copular verbs with a

discussion of the diﬁ‘erence of meaning between them. Kuno and Wongkhomthong

32

(1981) discuss the difference of meaning between pen and khu. Their assumption is that
the pen pattern is used for characterization and the khu pattern is used for identiﬁcation.
To illustrate, consider the following sentences;
(65) a. ruang thi chan cha bok khun pen/*khu ruang samkhan mak

thing that I will tell you is thing important very

“What I want to tell you is very important thing.”

[= Kuno and Wongkhomthong 1981, ex. 4a]

 

b. ruang thi chan cha bok khun k0 ‘pen/khu chan kam lang cha taeng ngan
thing that I will tell you is I about to will marry
“What I want to tell you is that I am going to marry.”
[= Kuno and Wongkhomthong 1981, ex. 4b]
(66) a. Ruang thi dichan bok khun John muwanni pen/*khu khwamching
thing that I tell Mr. John yesterday is fact
“What I told John yesterday is a fact.”
[= Kuno and Wongkhomthong 1981, ex. 5a]
b. Rung thi dichan bok khun John muwanni *pen/khu khwamching thi phom
thing that I tell Mr. John yesterday is fact that I
kamlung cha 100 hot khun fang
about to will tell give you listen
“What I told John yesterday is the fact that I am going to tell you.”
[= Kuno and Wongkhomthong 1981, ex. 5b]
(65a) and (66a) present one of the characteristics that their subject possesses. For

example, (65a) presents as a characteristic of what the speaker wants to tell the addressee

33

the fact that it is very important. (66a) characterizes what the speaker told John the day
before as factual. So they are characterizational sentences and the copular verb pen is
used. On the other hand, (65b) and (66b) are not characterizational sentences, but they
are identiﬁcational sentences. For example, (65b) identiﬁes what the speaker want to tell
the addressee as the fact that he is getting married. (66b) identiﬁes what the Speaker told
John the day before with the fact that he is going to tell the addressee. So khu is used as a
copular verb.
For cases in which pen and khu can be used interchangeably, the explanation has
to do with the speaker’s intention. Consider, for example, the following sentences:
(67) a. John pen=’khu khon thi chan rak
John is person that I love
“John is the person that I love.”
[=Kuno and Wongkhomthong 1981 , ex. 9b]
b. katoe pen/khu prathanathibadi khong saharatamerika
Carter is president of US
“Carter is the President of the United States of America.”
[=Kuno and Wongkhomthong 1981, ex. 10b]
(67a), for example, if the speaker’s intention is to present one of the characteristics that
John has, pen is used. On the other hand, if the speaker’s intention is to state that John
and the person that the speaker loves are one and the same person, khu is used.
Similarly, (67b) can be interpreted either as a sentence that presents one of Carter’s

characteristics, or a sentence which equates Carter and the President of the United States

34

 

of America. The characterizational copular pen is used for the former and the

identiﬁcational copular khu is used for the latter interpretation.

Note that for the sentences in (65a) and (66a) in which only the verb pen is
acceptable, when negated, their negation has to be mai (dai) pen6 and for the sentences
(65b) and (66b) in which only the verb khu is acceptable, when negated, their negation
has to be mai chai. This can be shown by the following.

(68) a. ruang thi chan cha bok khun mat dai pen/*mai chai ruang samkhan
thing that I will tell you not become be / not be thing important
mak
very
“What I want to tell you is not very important thing.”

b. ruang thi chan cha bok khun *mai dai pen/mat chai chan kam Iang
thing that I will tell you not become be / not be I about to
cha taeng ngan
will marry
“What I want to tell you is not that I am going to many.”

(69) a. Ruang thi dichan bok khun John muwanni mai dai pen/*mai chai
thing that I tell Mr.John yesterday not become be / not be
khwamching
fact

“What I told John yesterday is not a fact.”

35

b. Rung thi dichan bok khun John muwanni *mai dai pen/mat chai
thing that I tell Mr. John yesterday not become be / not be
khwamching thi phom kamlung cha lao hai khun fang
fact that I about to will tell give you listen
“What I told John yesterday is not the fact that I am going to tell you.”

As expected, in the case of (67a) if the speaker’s attention is to present one of the
characteristics that John does not have, the negation will be mai (dai) pen, as shown in
(703). On the other hand, if the speaker’s attention is to present that John and the person
that the speaker loves are not one and the same person, the negation will be mai chai, as
shown in (70b). Similarly, in (67b) if the speaker’s attention involves characterization,
mai (dai) pen is used for negation as shown in (71a) and if the speaker’s attention
involves identiﬁcation, mai chai is used for negation, as shown in (71b).

(70) a. John mat dai pen khan thi chan rak
John not become be person that I love
“John is not the person that I love.”

b. John mat chai khon thi chan rak
John not be person that I love
“John is not the person that I love.”

(71) a. katoe mat dai pen prathanathibadi khong saharatamerika
Carter not become be president of US

“Carter is not the President of the United States of America.”

 

’5 The word dai can be interpreted in many ways. Here the closest interpretation is
“become.” The discussion concerning the insertion of the word dot is beyond the scope

36

b. katoe mai chai prathanathibadi khong saharatamerika
Carter not be president of US
“Carter is not the President of the United States of America
So there is a clear difference between pen on the one hand, and khu and chai on the

other.

3.3.2 Restriction on Aspect: Comparing the Copular Verb in English with the

Copular verbs in Thai

Studies on aspect refer to studies on kinds of action, also known by the German
term Aktionsarten. Basically these studies involve a distinction between states and other
kinds of eventualities.

States are kinds of circumstances which do not have internal structure. At any
points of time, a state either holds or does not hold. The predicates expressing states
have the property of [+stativity], for example, like and know.

Events, on the other hand, are kinds of circumstances which have internal
structure and which culminates at a certain point of time. The predicates expressing
events are therefore, [-stativity] predicates such as read and go.

The distinction between states and events involve subject control. Generally
states are not under the control of the subject. For example, to know or not to know is not
under one’s control. Events, on the other hand, are under the control of subject. For

example, to go or not to go is under one’s control.

 

of this paper.

37

Usually copular verbs are considered to be [+stativity] verbs. The copular verb in
English and the copular verbs in Thai, though both are considered [+stativity], show
different restrictions on aspect, for example, the restriction on adverbial modiﬁcation and

modal verbs.

3.3.2.1 Restriction on Adverbial Modiﬁcation

 

Restriction on adverbialjhrases indicating termination

Adverb phrases such as in X years are considered as indicating termination, while
adverb phrases such as for X years do not indicate termination. To illustrate, consider the
following examples.

(72) a. John built a house in 2 years

b. *John walked in 2 years.

(73) a. *John built a house for 2 years.

b. John walked for 2 years.

The predicate built a house suggests termination, but the predicate walked does not. So
the predicate built a house can occur with the adverb in 2 years but cannot occur with the
adverb for 2 years. On the contrary, the predicate walked cannot occur with the adverb
in 2 years and can occur with the adverb for 2 years.

Both the verbs build and walk are [-stativity] verbs. Generally these adverb

phrases are associated with [~stativity] verbs. Copular verbs are considered to be

38

[+stativity] verbs. However, the verb be in English can occur with these adverb phrases,
as shown by the following.
(74) a. John will be a doctor in 2 years.

b. John has been a doctor for 2 years.
However, the adverb phrase in 2 years in (74a) does not seem to express termination but
beginning. In particular, what described by (74a) is that in the next two years, John will
start performing a doctor. The adverb phrase for 2 years in (74b) seems to be allowed
because the verb phrase be a doctor expresses a kind of event. In particular, to become
or not to become a doctor seems to be under one’s control. So it has some property of
events.

The copular verbs in Thai Show some restrictions. Only the copular verb pen can
occur with the adverb phrase in X years and for X years. The copular verbs khu and chat
cannot.

(75) a. John cha pen hmo nat' wela ik song pi
John FUTURE be doctor in time more two year
“John will be a doctor in two years.”

b. John pen hmo ma pen wela song pi

John be doctor PERFECT time two years
“John has been a doctor for two years.”

(76) a. *John cha khu hmo nai wela ik song pi
John FUTURE be doctor in time more two year

“John will be a doctor in two years.”

39

b. *John khu hmo ma pen wela song pi
John be doctor PERFECT time two years
“John has been a doctor for two years.”
(77) a. *John cha mat chai hmo nai wela ik song pi
John FUTURE not be doctor in time more two year
“John will not be a doctor in two years.”
b. *John mai chai hmo ma pen wela song pi
John not be doctor PERFECT time two years
“John has not been a doctor for two years.”
So it seems that the copular verbs that allow adverbial modiﬁcations like in X time and
for X time, carry an eventive interpretation. In particular, they carry the sense of
performing. The verb be when combined with a predicate nominal like a doctor, and the

verb pen, therefore, carries this sense while the verbs khu and chai do not.

Restriction orgdverbﬁial phrase indicating a mrticuliar point of time

An adverb phrase such as “when I was 27 years old” indicates a particular point
of time. In English this can modify the copular verb be, as shown by (78).
(78) John was a doctor when he was 27 years old.

In Thai, the adverb phrase ton ayu 27 pi “when he was 27 years old” can modify

only the verb pen, not the verbs khu and chai, shown by the following.

40

(79)

(30)

(81)

John pen hmo ton ayu 27 pi

John be doctor at the time age 27 year

“John was a doctor when he was 27 years old.”
*John khu hmo ton ayu 27 pi

John be doctor at the time age 27 year

“John was a doctor when he was 27 years old.”
*John mat chai hmo ton ayu 27 pi
John not be doctor at the time age 27 year

“John was not a doctor when he was 27 years old.”

Adverb phrases like in the past do not seem to indicate a particular point in time.

This kind of adverbs can occur with the copular verb be in English and all the copular

verbs in Thai as shown below.

(82)

(83)

(34)

Mua kon John pen hmo
When before John be doctor
“In the past John was a doctor.”
Mua kon John khu hmo

When before John be doctor

“In the past John was a doctor.”
Mua kon John mat chai hmo
When before John not be doctor
“In the past John was not a doctor.”

The adverb phrase indicating a particular point of time seems to involve the idea

cf transition. As the verb phrase (John) was a doctor is modiﬁed by the adverb phrase

41

when he was 27 years old, it involves the idea that before John was 27 years old, he was
not a doctor, but after he turned to 27 years old he became a doctor.
An adverb phrase that does not indicate a particular point of time, such as in the
past, on the other hand, does not involve any idea of transition.
So it seems that the verbs that allow the adverbial modiﬁcation indicating a
particular point of time are associated with the interpretation of transition. The verb be
and pen can be, therefore, associated with the idea of transition while the verb khu and

chai are not.

3.3.2.2 Restriction on Modal Verbs

Restriction on the modal verb expressiggabilitv and permission

This modal verb seems to be of particular interest. The modal verb can is used to
express ability, and permission. In some cases it can occur with the verb be, as shown by
the following.

(85) a. John can be a doctor

b. *John can be tall
Can can occur in (85a) but not in (85b). This can be accounted for by the fact that the
predicate be tall is carrying a higher degree of stativity than the predicate be a doctor; in
particular, being tall is permanent and not under the control of the subject but being a

doctor is not permanent and it is under the control of the subject. Therefore while the

42

predicate be tall cannot occur with the modal verb expressing ability and/or permission,
the predicate be a doctor can.
In Thai the modal verb dai “can” can occur with the copular verb pen but it
cannot occur with either the copular verbs khu or chai, as shown below.
(86) John pen prathan dai
John be president can
“John can be a president.”
(87) *John khu prathan dat'
John be president can
“John can be a president.”
(88) *John chai prathan mat dai
John be president not can
“John cannot be a president.”
So it seems that the verbs khu and chai cany a higher degree of stativity than the verb

pen.

Restriction on guides expressin ro essive form

In English, the verb to be cannot occur in the progressive form, as shown by the
following.
(89) I"John is being a doctor

In Thai the progressive form is expressed by the word yu. The verb pen can occur

in the progressive form but the verbs khu and chai cannot, as shown by the following.

43

(90) John pen hmo yu
John be doctor PROGRESSIVE
“John is being a doctor.”
(91) *John khu hmo yu
John be doctor PROGRESSIVE
“John is being a doctor.”
(92) *John mai chai hmo yu
John not be doctor PROGRESSIVE
“John is not being a doctor.”
Generally the progressive form can only occur with [-stativity] verbs. However, it
can occur with the verb pen. So it seems that the verb pen, compared with khu and chai

and with English be, is the most likely to behave like a [-stativity] verb.

Restriction on particles expressing tenses

Languages have different ways to express tenses. English has tense-inﬂections.
Thai, on the other hand, express tenses by particles such as cha (future tense) and khu
(past tense). Without the particles, the tense of the verb is understood as present tense.

The copular verb be in English can occur in any tense, in particular, present,

future and past tenses.

(93) a. John is a doctor.

b. John will be a doctor.

c. John was a doctor.

The copular verbs in Thai can occur in present tense. However, only the verb pen
can occur with the particles expressing future and past tenses. The verb khu and chai
cannot.

(94) a. John pen hmo
John be doctor
“John is a doctor.”
b. John cha pen hmo
John FUTURE be doctor
“John will be a doctor.”
c. John khoey pen hmo
John PAST be doctor
“John used to be a doctor.”
(95) a. John khu hmo
John be doctor
“John is a doctor.”
b. *John cha khu hmo
John FUTURE be doctor

“John will be a doctor.”

45

c. *John khoey khu hmo
John PAST be doctor
“John was a doctor.”
(96) a. John mai chai hmo
John not is doctor
“John is not a doctor.”
b. *John cha mai chai hmo
John FUTURE not be doctor
“John will not be a doctor.”
0. *John mat khoey chai hmo
John not PAST be doctor
“John was not a doctor.”
It is worth noting here the fact that generally [-stativity] verbs cannot occur in the simple
present tense. They have to be either in the future or in the past tense.
(97) a. *John builds a house
b. John will build a house
c. John built a house
To express the present tense, [-stativity] verbs have to be in the present
progressive.
(98) John is building a house.
The copular verbs khu and chai in Thai behave totally differently. They have to be in the

present tense and as shown in the previous section, they cannot be in progressive form.

So it seems that comparing to the verbs be and pen, the verbs khu and chai have a higher
degree of [+stativity].

In summary all of the copular verbs, in particular be pen khu and chai express
states but only be and pen allow eventive interpretation which is the sense of performing
and transition and only pen allows eventive form which is progressive form.

So the copular verb in English and the copular verbs in Thai are all different in
their aspectuality. While English has only one copular verb (be) which has a medium
degree of stativity, Thai has three copular verbs-two (khu and chai) expressing a high

degree of stativity and one (pen) expressing a low degree of stativity.

3.3.3 Restriction on Aspect: Comparing the Copular Verbs with Other Stative

Verbs in Thai

While the polarity sensitive copular verbs in Thai show many restrictrions on

aspect, in particular restrictions on adverbial modiﬁcation and modal verbs, other

[+stativity] verbs do not. Consider, as representatives, the [+stativity] verbs ru “know”

and chop “like” in the following examples.

3.3.3.1 Restriction on Adverb Phrases Indicating a Particular Point of Time

While the polarity sensitive copular verbs cannot occur with an adverb phrase

indicating a particular point of time, Stative verbs such as ru “know” and chop “like” can.

47

(99) John ru phasathai ton ayu 27 pi
John know language Thai at the time age 27 year
“John knew Thai when he was 27 years old.”
(100) John chop Mary ton ayu 27 pi
John like Mary at the time age 27 year
“John liked Mary when he was 27 pi
As the adverb phrase like ton ayu 27 pi “when he was 27 years old” involves the idea of
transition, it seems that only the polarity sensitive copular verbs do not allow this idea

while the other [+stativity] verbs do.

3.3.3.2 Restriction on Particles Expressing Tenses

While the polarity sensitive copular verbs cannot occur with the particles
expressing future and past tenses, the [+stativity] verbs ru “know” and chop “like,” can,
as shown by the following.

(101) a. John cha ru phasathai hlangchak chop karn oprom ni
John FUTURE know Thai aﬁer ﬁnish training this
“John will know French after ﬁnishing this training.”
b. John khoey ru phasathai tae tonni khao lum laew
John PAST know Thai but now he forget PERFECT

“John knew Thai but now he forgot it.”

48

(102) a. John cha chop Mary hlangchak khao dai hen thoe
John FUTURE like Mary aﬁer he get see her
“John will like Mary after he sees her.”
b. John khoey chop Mary tae ton ni khao chop Susan
John PAST like Mary but now he like Susan
“John liked Mary but now he likes Susan.”
So it seems that only the polarity sensitive copular verbs are limited to the present tense
form. The other [+stativity] do not show this limitation.

It is worth noting here that the verbs ru and chop in Thai behave in the same way
as the verbs know and like in English. In particular, they can occur with the adverb
phrase indicating a particular point of time and they can be in present tense, future tense,
and past tense.

So it seems the polarity sensitive copular verbs have peculiar aspectual facts. The
other verbs of the same kind do not show the aspectual restrictions that the polarity

sensitive copular verbs do.

3.3.4 Summary of the Discussion on Aspect

So far it seems that the polarity sensitive copular verbs in Thai possess a kind of
peculiarity. The aspectual distribution of the copular verbs in Thai differs from the
copular verb be in English and other [+stativity] verbs.

In section 3.3.2, I compared the copular verbs in Thai with the copular verb in

English. It seems that the verb pen has a lesser degree of [+stativity] than the verb be in

49

English and the verbs khu and chai have a greater degree of [+stativity] than the verb be
in English, as shown by the following.

The verb pen in Thai
0 expressing states
0 eventive interpretation (performing and transition)
0 progressive form

The verb be in English
0 expressing states
0 eventive interpretation (performing and transition)
0 “progressive form

The verb khu and verb chai in Thai
0 expressing states
0 *eventive interpretation (performing and transition)
0 I"progressive form

This assumption seems to be compatible with what was discussed by Kuno and
Wongkhomthong (1981). Since the copular verb pen is more eventive than khu, pen is
used for characterization, which can be changed over time and the copular verbs khu and
chai are used for identiﬁcation, which can hardly be changed over time.

I show in Section 3.3.3 that other [+stativity] verbs do not show the restrictions on
aspect that the polarity sensitive copular verbs do.

So the aspectual distribution of the polarity sensitive copular verbs differs from
both that of the copular verb be in English and other [+stativity] verbs in Thai. Crucially

the copular verb be in English and the other [+stativity] verbs in Thai are not polarity

50

sensitive. So it seems possible that the aspectual fact might be related to the polarity
sensitivity facts.

Moreover, both aspect and polarity sensitivity are the restrictions on a sentence
level. In particular, both of them have to be interpreted with respect to a verb phrase in a
sentence. So it seems possible that there might be a connection between the aspectual

properties of Thai copular verbs and their property of polarity sensitivity.

3.4 An Alternative Approach to the Copular Verbs in Thai

While the present paper states that Thai has 3 copular verbs: pen, khu and chai,
Prof. Lockwood, offers another approach to the copular verbs in Thai. He suggests that
the copular verbs in Thai might be divided into 2 morphemes. One is pen and another is
a morpheme which consists of 2 allomorphs, in particular khu and chai.

Supporting this, khu and chai share the same meaning and are in complementary
distribution, in particular, one is for positive contexts and one is for negative contexts, as
discussed in section 3.3.1.

As khu and chai have no etymological relation, they are considered to be
suppletive allomorphs. By treating the verbs khu and chai as allomorphs of the same
morpheme will correspond to English verbs, such as go and went which are allomorphs
of the same morpheme. The two allomorphs khu and chai would share the same
meaning but one is PPI and one is NPI. The two allomorphs go and went would share the

same meaning but one is for present tense and one is for past tense.

51

However, this approach does not bring about explanation for the polarity
sensitivity, in particular why one has to be PPI and one has to be NPI. While it is
common in English that every verb has a present tense form and past tense form, it is
uncommon for a verb to have one form for positive contexts and one form of negative
contexts.

This approach does not either explain the aspectual facts, in particular, why this
single morpheme (the morpheme that consists of 2 allomorphs khu and chai) has the
aspectual restriction that the other morphemes do not.

Therefore, it is another approach to the copular verbs in Thai. However, the

fundamental problems concerning the polarity sensitivity and aspectuality still remain.

3.5 Conclusion

In this chapter 1 have shown that the polarity sensitive copular verbs in Thai pose
problems to the theory of polarity sensitivity. However, rather than relating to the
licensing issue, the data from the polarity sensitive copular verbs in Thai presents a big
problem for the sensitivity issue.

Accounts focusing on the licensing problem, in particular Ladusaw
(1980,1982,I983) and Linebarger (1981,1987,1989), are able to account for NPI
licensing in Thai.

Israel’s (1996) account which is focusing on the sensitivity problem cannot
account for the polarity sensitive copular verbs in Thai. His explanation relates the

‘ semantic features that the polarity sensitive items carry to the occurrences of the polarity

52

sensitivity items. As copular verbs carry no semantic meaning, his explanation does not
apply to the polarity sensitive copular verbs in Thai.

Crucially the polarity sensitive copular verbs in Thai show restrictions on aspect.
By comparing with other verbs of the same kind, the polarity sensitive copular verbs do
not share any common aspectual distributions with the other verbs. So it might be
possible that the property of polarity sensitivity that the copular verbs possess is related

to their aspectual restrictions.

53

CHAPTER 4

CONCLUSION

This thesis has investigated the restriction on the distribution on the polarity
sensitive copular verbs in Thai. Earlier studies on polarity sensitivity cannot account for
the case of polarity sensitive copular verbs in Thai.

There seem to be two main questions for the studies on polarity sensitivity: the
licensing question and the sensitivity question. While the licensing question deals with
makes certain contexts license polarity sensitivity the sensitivity question involves what
makes certain forms sensitive to these contexts.

Much of the previous research has focused on the licensing question. However, it
is the sensitivity question, rather than the licensing question, that presents a big problem
in Thai.

Since generally the NPI in Thai is generally licensed by the negation which is the
simplest case of NPI licensing, any accounts on NPI licensing, in particular, Ladusaw
(1980,1982,I983) and Linebarger (1981,1987,1989) seem to apply.

The sensitivity seems to be a big problem in Thai. The previous research
focusing on the sensitivity, in particular Israel (1996) fail to account for the polarity
sensitive copular verbs in Thai.

Israel (1996) claims that PSIs are speciﬁed for two semantic features: quantitative
value (q-value) and informative value (i-value), and the interaction of these two features
makes them sensitive to certain contexts. Obviously, his explanation relates the semantic

features that the PSIs carry to the distribution of P815. As the copular verbs carry no

54

semantic meaning, their distribution cannot be explained by the semantic features. His
account, therefore, is not working for the case of polarity sensitive copular verbs in Thai.

I have shown that the polarity sensitive copular verbs in Thai possess a kind of
peculiarity in terms of aspect. The other verbs of the same type, in particular, the verb
pen in Thai, the verb be in English and the other [+stativity] verbs, such as ru “know”
and chop “like,” do not Show the aspectual restrictions that the copular verbs khu (PPI)
and chai (NPI) do and at the same time they are not polarity sensitive. As aspect and
polarity sensitivity are both restrictions over a verb phrase, in particular, they have to be
interpreted with respect to a verb phrase in a sentence, it seems possible that there might
be connection between the aspect and the polarity sensitivity. So the aspectual facts
might be related to the polarity sensitivity of the copular verbs.

The problem remaining is to ﬁnd a relationship between the polarity sensitivity of
khu and chai and their aspectual constraints. There seem to be three possible areas in
which this relationship could be found. The relation between aspect and polarity
sensitivity might involve three possible kinds of facts: semantic facts, syntactic facts and
pragmatic facts.

Semantic facts need to be considered because the copular verbs in Thai carry
some semantic features. Generally copular verbs carry no semantic meaning. However,
as Kuno and Wongkhomthong (1981) showed, one of the copular verbs is used for
characterization and one is used for identiﬁcation. This means that logically the copular
verbs must carry some semantic features. To explain the aspect and polarity sensitivity,

this fact cannot be ruled out.

55

The syntactic facts I have in mind have to do with hierarchical structure. In a tree
structure the nodes TP (tense phrase) and NEGP (negation phrase) are higher than the
node VP (verb phrase). There might be some reasons that a VP containing khu or chai
cannot be in a position under the TP and in some ways the VP khu prevents the NEGP
but the VP chai requires the NEGP. It seems possible that the structural explanation for
the aspect and polarity sensitivity is on this track.

A pragmatic fact that might be relevant is the fact that the word chai is
ambiguous. In particular, the word chai is a copular verb, similar to “be” in English and
also a word used to show that a statement is correct or that the speaker agrees, similar to
“Yes” in English. This ambiguity might affect the use of aspect and specify the polarity
of the word. This fact, therefore, should be in consideration as well.

So although the explanation of how the aspect and polarity sensitivity are related
has not been offered, the possible directions in which they might be related have been
pointed out.

The polarity sensitivity is considered one of the mysteries in linguistic theory.
The data from the polarity sensitivity copular verbs in Thai has shown another way that
the polarity sensitive items might vary. The earlier theory of polarity sensitivity has not
resulted in an adequate explanation for these polarity sensitivity items. It is hoped that

this thesis can be at least motivation for the future research on polarity sensitivity.

56

BIBLIOGRAPHY

Fauconnier, G. (1975) Polarity and the scale principle. In R.E. Grossman, L.J. San and
T.J. Vance (eds), Proceedings of the 11'” Meeting of the Chicago Linguistics
Society, 188-99.

Israel, M. (1996) Polarity Sensitivity as Lexical Semantics, Linguistics and Philosophy 5,
619-666.

Klima, E. (1964) Negation in English. In J .A. Foder and J. Katz (eds), The structure of
Language, 246-323. New York: Prentice-Hall.

Kuno, S. and Wongkhomthong, W. (1981) Characterizational and Identiﬁcational
Sentences in Thai. Studies in Language 5, 65-109.

Ladusaw, WA. (1980) Polarity Sensitivity as Inherent Scope Relation. Garland, New
York.

------- --(1982) On the Notion of Affective in the Analysis of Negative-polarity Items.
Journal of Linguistics Research 2, 1-16.

--------- -(1983) Logical Form and Conditions on Grammaticality. Linguistics and
Philosophy 6, 373-392.

------- -—(1996) Negative Polarity Items. In S. Lappin (ed.) The Handbook of
Contemporary Semantic Theory, 321-341.

Linebarger, MC. (1981) The Grammar of Negative Polarity. Indiana University
Linguistics Club, Bloomington.

--------- -(1987) Negative Polarity and grammatical representation Linguistics and
Philosophy, 10, 523-87.

--------- -(l980) The Grammar of Negative Polarity, Ph.D. Dissertaion, MIT, Cambridge,
Massachusetts.

Progovac, L. (1994) Negative and Positive Polarity: A Binding Approach. Cambridge;
University Press.

--------- -(1992). Negative polarity: A Semantico-syntactic approach. Lingual 86: 271-299.

Royal Institute (1982) Romanization Guide for Thai Script. Prayurawong Company
Limited, Bangkok.

57

Uppakitsinlapasam, Phaya (1964) Waiyakomthai [Thai Grammar], Thai Wattana Phanit,
Bangkok.

Van der Wouden, T. (1994). Polarity and Illogical Negation. In Kanazawa, M., and
Pinon, C .(eds), Dynamics, polarity and quantiﬁcation, 17-45.

58

       

HICHIGRN STRTE UNIV. LIBRQRIES
llllllllllllll lllllllllllllllllllllllllllllllllllllll
312 3017700679

|
9