WIlMlWlNllHWWWIHWHIHHWUWI

      
   

CO
(DLOOO

means

00.5100

 

"$222!“ IIHHMIWujluugmmlwumm
University

 

 

This is to certify that the

thesis entitled

JAPANESE FILLERS AND PSYCHOLOGICAL DISTANCE

presented by

Yoshiko Yamamoto

has been accepted towards fulﬁllment
of the requirements for

M.A. degree in Linguistics

 

 

2m Lot/AF!“—

Major professor

Date 12/ 'J 0' if

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

 

 

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

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1m W14

 

JAPANESE FILLERS AND PSYCHOLOGICAL DISTANCE
By

Yoshiko Yamamoto

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 African Languages

1998

ABSTRACT
JAPANESE FILLERS AND PSYCHOLOGICAL DISTANCE
By

Yoshiko Yamamoto

The present study investigates a relation between psychological distance between
speakers and the use of fillers, as well as the functions of the two most commonly
used ﬁllers, nan/ea ‘like’ and and ‘well’, in Japanese conversation. Based on the
data Of twelve naturally-occurring conversations, the following conclusions are
made: Fillers are categorized into language-production—based type and socially
motivated type; some words cannot be categorized either as fillers or words in other
categories; and some fillers cannot be clearly categorized into language-production-
based type or socially motivated type. Speech styles affect both the frequency and
the kind of fillers. More fillers occur in the polite style than in the plain style.
The casual sounding ﬁller nan/ea ‘like’ occurred more frequently in the plain type.
Relatively formal sounding ﬁllers, such as 4720 ‘well’ and 50720 ‘well’, on the other
hand, occurred more frequently in the polite type. This supports the claim that
more fillers occur when the Speaker feels psychologically distant towards the
addressee.

Fillers, especially have not received much attention in the past, since they
are simply regarded as something that just fills potential silence. However, the
present study shows that they play important roles in structuring discourse

coherence in much the same way as other discourse markers.

Copyright by
YOSHIKO YAMAMOTO
1998

To my brother

iv

ACKNOWLEDGMENTS

My deepest gratitude goes to Dr. Mutsuko Endo Hudson, the chair of my
thesis committee. Without her insightful comments, suggestions and patience, I
would not have been able to ﬁnish this thesis. I also would like to thank Dr.
Barbara Abbott and Dr. Denise Troutman for their valuable comments and

suggestions.

List of Tables

Chapter 1

Chapter 2

Chapter 3

Chapter 4

TABLE OF CONTENTS

Introduction

1.1. Phenomena to be investigated

1.2. Hypothesis

1.3. Framework

1.4. Preliminaries
1.4.1. Demonstratives ano, 50120 and kono
1.4.2. Other category items
1.4.3. Speech style in Japanese

Review of related literature
2.1. Definitions of ‘discourse markers’ and ‘fillers’
2.2. Previous studies on discourse markers
2.2.1. Sacks, Schegloff and Jefferson(1974)
2.2.2. Owen (1981)
2.2.3. Shourup (1982)
2.2.4. Stubbs (1983)
2.2.5. Shiffrin (1987)
2.2.6. Holms (1986)
2.2.7. Bloch (1996)
2.3. Previous studies on fillers in Japanese
2.3.1. Maynard (1989)
2.3.2. Saito (1992)
2.3.3. Cook (1993)
2.3.4. Philips (1998)
2.4. Summary

Methods and procedures
3.1. Data
3.2. Participants

3.3. Procedures
3.4. Methods of analysis

Results and analysis

4.1. Total number and the rate

4.2. Fillers and speech styles

4.3. Fillers and questionnaire results
4.3.1. Fillers and relationship between
interlocutors

vi

viii

\OOOVCb-hwi—tv—s

Chapter 5
Appendix A
Appendix B

Appendix C

Bibliography

4.3.2. Fillers and impressions of conversation
and of interlocutor

4.4. Environments and functions of fillers
4.4.1. Environments of language-production-

based fillers

4.4.2. Functions of language-production-based
fillers
4.4.3. Environments of socially motivated ﬁllers
4.4.4. Functions of the filler nan/ea ‘like’
4.4.5. Functions of the filler ano ‘well’

4.5. Summary

Conclusion

vii

35
37

37
40
40
4o
44
47
49

51

52

53

54

LIST OF TABLES

Table 1 Summary of am, sono and kono

Table 2 Summary of other category items

Table 3 Summary of -des- and -mas-the polite Style

Table 4 Participants

Table 5 Summary of ﬁller

Table 6 Language-production-based type 8: socially motivated type
Table 7 Number of occurrences of each type of filler

Table 8 Occurrences of -des- and -mas- in each conversation

Table 9 Occurrences of fillers in each speech style

Table 10 Summary of kinds of fillers for each speech style

Table 11 Fillers and relationship between interlocutors

Table 12 Fillers and impressions of conversation and of interlocutor
Table 13 Summary of environments of language-production-based fillers
Table 14 Summary of nan/ea ‘like’ and mm ‘well’ with position
Table 15 Summary of nan/ea ‘like’ on Shiffrin’s five planes of talk

Table 16 Summary of 4720 ‘well’ on Shiffrin’s ﬁve planes of talk

viii

23

25

28

29

31

32

33

35

37

38

4O

44

47

Chapter 1
Introduction
1.1. Phenomena to be investigated

The present study investigates a possible relation between psychological
distance between speakers and the use of fillers as well as the function of fillers in
Japanese dyadic conversation. Fillers have not received much attention in the past,
since they are regarded as something that just fills potential silence. However, in the
present study they are regarded not linguistically meaningless, rather they play
important roles structuring discourse coherence in much the way as other discourse
markers do.

Fillers are defined by Maynard (1989:30) as ‘a broad range of utterances that do
not carry identifiable or relevant propositional meanings.’ In this study the following
expressions are regarded as ﬁllers: 4220(0) ‘well’, sono(o) ‘well’, kono(o) ‘well’, nan/ea
‘like’, not ‘uh, well’, iya(a) ‘well’, mad ‘kind Of’, ee(to) ‘uh, well’, uun(to) ‘uh, well’.
Many fillers have characteristics of ‘discourse markers’, which according to Schiffrin
(1987:328) ‘bracket units of talk, are syntactically detachable, are used in initial position
of utterance, have prosodic contours, and have no or vague meaning.’ She further
defines discourse markers as ‘contextual coordinates,’ which index an utterance to the
local contexts in which utterances are produced and in which they are to be
interpreted.’ (p. 326) Following Philips (1998), fillers are regarded as one type of
discourse marker in Japanese in this study along with certain conjunctives (e.g. sorede

‘and’) and interactional particles (e.g. ne ‘TAG QUESTION’). Detailed definitions of

fillers and discourse markers are provided in section 2.1.

The purpose of this study is to provide analysis to Show that fillers have
important functions as a content softener showing hesitation and/ or uncertainty not to
be imposing to the addressee, as an attention getter evoking the addressee’s cooperation,
and as a softener to reduce an abruptness of the upcoming content of the utterance.
Fillers also function to signal a repair and intention to continue. It is interpreted from
these functions that the fillers function to help create maximum agreeableness between
the speakers as pointed out by Maynard (1989). In this sense, fillers contribute to
discourse coherence.

Example (1) illustrates the use of the ﬁllers, mm ‘uh’, 44 ‘oh’, nan/ea ‘like’, sono

‘well’, and maa ‘kind of’ all in one utterance.

(1) R: a. mm a sooyuu imi de u-watashi mo sa chuugaku de nanka
F F that meaning I too 1P junior high in F

b. sono maa shukudai toka o kanarazu kake tte in fun ni itteru

F F homework like DO surely write QT tell
c. no ne
NOM IP

‘Well, in that sense. In the junior high, I too tell (the students), well, to write
down the homework.’

Although there have been a number of studies on discourse markers in both
English (e.g. Fraser 1998, Schiffrin 1987, Schourup 1982) and Japanese (e.g. Hudson

1998, Ito 1995, Maynard 1989, Saito 1992, Philips 1998), the functions of Japanese

fillers have yet to be thoroughly explored. Only a few studies have been carried out

on the use of English fillers (e.g. Owen 1981, Stubbs 1983) and even fewer in Japanese
(Cook 1993). Cook (1993), for example, focuses only on the functions of the ﬁller,
4720 ‘well’, and there have been no studies on the possible relation between
psychological distance and the use of Japanese fillers. The present study examines
twelve naturally-occurring dyadic conversations in Japanese both quantitatively and
qualitatively. It is hoped that the findings in the present study will, in general, be
useful in future research on Japanese fillers and discourse analysis. This study has two
goals: (1) to identify the relation between psychological distance and the use of fillers
and (2) to identify the functions of each filler in Japanese.

The organization of this thesis is as follows. The remainder of Chapter 1
outlines the hypotheses to be tested and the theoretical framework adopted. Chapter
2 summarizes previous studies on discourse markers with special attention to fillers.
Chapter 3 outlines the methods and procedures of the present study. Chapter 4
discusses results and analysis. Chapter 5 offers conclusions.

1.2. Hypothesis

Based on Philips (1998) on Japanese discourse markers and also on my pilot
study (Y amamoto 1997), the following hypothesis has been formed:

Hypothesis: There is a relation between psychological distance and the use of

fillers; i.e. the more the speaker feels psychologically distant toward the

addressee, the more frequently fillers are used.
In Philips’s study, the total rate of filler marker use were much higher in formal

conversation, in which the ma form of verbs (polite verb ending) and the desu form

(polite form of the copula and polite ending of i- type adjective) are used, than in the

informal type in which the plain forms of verbs, adjectives and the copula are used. In
the pilot study, the number of occurrences of sentence-ﬁnal desu ‘be’, masu ‘do’, and
leeigo (polite language) expressions, which are used when the psychological distance
between the speakers is distant, was compared with the number of occurrences of the
fillers 4120(0) and $0n0(0) ‘well’. The results showed that the number of occurrences of
the fillers was greater when the distance greater. The present study employs more
fillers than just 4720(0) and 50n0(0) ‘well’ with larger data and additional methods to
measure psychological distance.
1.3. Framework

The framework adopted in the present study is that of discourse analysis

developed by Schiffrin (1987) in which key assumptions in discourse analysis are listed

as follows (p. 3):

1. Language always occurs in a context.
2. Language is context sensitive.

3. Language is always communicative.
4. Language is designed for communication.

The properties of discourse are further listed as follows (p. 6):

1. Discourse forms structures.

2. Discourse conveys meanings.
3. Discourse accomplishes actions.

She states that no one of these properties can be understood without attention to the
others (p. 13). Thus, all of these properties contribute to the overall sense —to the
coherence— of discourse.

Schiffrin’s framework analyzing English discourse markers was adopted in this

study of Japanese ﬁllers, which are regard as a type of discourse marker. Fillers are
regarded as discourse markers for the following reasons: First of all, fillers satisfy all the
conditions for discourse markers (Schiffrin 1987:328): discourse markers bracket units
of talk, are syntactically detachable, are used in initial position of utterance, have
prosodic contours, and have no or vague meaning (see Section 2.1 for detailed
discussion of these conditions). Secondly, the data indicates that many fillers show
speaker’s orientation toward both the speaker and the bearer, toward the proposition
and discourse context in which the conversation is taking place. Thus, fillers can be
regarded as ‘contextual coordinates’ as deﬁned by Shiffrin which provides coordinates
to the contexts (p.326). Therefore, fillers are regarded as discourse markers in the
present study.

Schiffrin analyzes eleven English discourse markers as contextual coordinates

which operate on the following ﬁve different planes of talk (pp. 24-29):

1. Ideational Structure — deals with propositions themselves, their cohesive

relations, topic relations, and functional relations.

2. Action Structure - reﬂects the sequence of speech acts, such as

requesting, warning, and asking.
3. Exchange Structure — deals with turn-taking.

4. Participation Framework - deals with speaker-utterance relationships as

well as speaker-bearer relationships.

5. Information State - focuses on cognitive capacities, including the
organization and management of knowledge (what a speaker knows, what a

bearer knows) and meta-knowledge (what speakers and hearer know about

each other’s knowledge).

According to Schiffrin, all the discourse markers are used on more than one plane
simultaneously, but on one plane primarily. For example, the primary function of 0b
is to mark information state transitions like recognize, receive, and re—evaluate
information. On the other hand, 019 also works in action structure since it marks
certain action like asking for clarifications. Well has its primary function in the
participation framework because it anchors a speaker into an interaction as a
respondent. It also functions in information states, idea structures, and action
structures.

Schiffrin’s comprehensive approach is appropriate for this study since Japanese
fillers cannot be analyzed without considering various elements of talk, such as speech
acts, text cohesion, and turn taking, and they seem to indeed function on different
discourse planes. By adopting Schiffrin’s framework, the multifunctionality of ﬁllers
can be accounted for as was pointed out in Philips (1998). This study of Japanese
fillers focuses primarily on the participation framework, since it is relevant tO
psychological distance between speakers.

1 .4. Preliminaries

Many fillers are originally lexical items that belong to categories such as
demonstrative adjectives (e.g. 4720 ‘that’) and indeﬁnite noun (e.g. nan/e4 ‘something’).
According to Schourup (1982), a lexical item becomes a filler when it is used routinely,
losing its original meaning. In analyzing the functions of Japanese ﬁllers, ﬁller

identification is often problematic because the distinction between an adjective, for

example, and a filler is not clear—cut. Shourup (1982:10—13) states: ‘many such items
have come to be so closely associated with particular discourse situations that they may
be considered conventional responses to these situations... The meaning of items of
verbal routine is thought to differ from that of other, nonroutine items in that the
literal meaning may be outshown’ (p.10). In the next two subsections words that can
be categorized as both ﬁllers and other lexical items are discussed.
1.4.1. Demonstratives 4220, sono and kono

A220, sono and leono are originally demonstrative adjectives in Japanese preﬁxed
to a noun. A220 means ‘that (far away from both the speaker and the hearer)’, $0220
means ‘that (near the hearer)’ and leono means ‘this’. As demonstrative adjectives,
they may point to an object (in deictic use) or refer to something previously mentioned
or implied in the context (in anaphoric use). When these words are used as fillers, the
last vowels are often prolonged, as in 42200, 502200 and 1202200, respectively, but they are
never prolonged in their use as demonstrative adjectives. Table 1 summarizes the
differences between the demonstrative adjectives and the ﬁllers 4220, 50220 and kono.

Table 1 Summary of 4220, sono and leono

 

as demonstratives as fillers

 

4220 4220 + NP ‘that (far away from the 4220(0) ‘well’
speaker and the hearer)’

 

50220 sono + NP ‘that (near the hearer)’ 50220(0) ‘well’

 

 

 

 

 

k0220 kono + NP ‘this’ k0n0(0) ‘well’

 

The data shows some occurrences of 4220 and sono which cannot be clearly

categorized into either demonstrative adjectives or fillers due to the fact that it is

 

impossible to ascertain the two speakers’ background knowledge. For example, some
occurrences of 4220 can be interpreted as meaning ‘that which we know in common’ or
just ‘well’. This issue will be discussed in more detail in section 4.2.
1.4.2. Other category items

N422/e4 is originally an indeﬁnite noun meaning ‘something, anything’, but as a
ﬁller it means ‘like’. A4, iya, cc and mm are all originally interjections, but also
function as fillers. Table 2 summarizes the meanings of these words.

of other items
of speech as lexical item

Table 2 S

     

   

noun mething, anything ' kind of
well

' terjection well

on well
' 11 well

To categorize these words as either ﬁllers or words in other categories, the context
(time, place and what was previously discussed, among others), the turns and the accent
and intonation of the word must be taken into consideration. In addition, the
speaker’s current information state and prior experience must be known. In order to
categorize these words, the researcher must solicit information from the participants
about their experience at the time of recording, which is difﬁcult to achieve, if not
impossible. Thus, it is important to recognize that there are some occurrences of
words that cannot be categorized as either ﬁllers or words in other categories. This

issue will be discussed in more detail in section 4.2.

1.4.3. Speech style in Japanese

In Japanese there is a distinction between polite (or formal) and plain (or
informal) Speech. In the polite style, -m45- (the polite suffix attached to verbs) and
-des~ (the polite suffix attached to the copula and to i~ type adjective) are used. Makino

and Tsutsui (1986:42) state:

the formal style is normally used when one is NOT speaking intimately or
personally with someone who belongs to his in-group. The informal style is

used when one is speaking with one’s own in-group... The formal style is
marked by -m45— or -des-

Table 3 summarizes inﬂections of -des- and -m45- in the polite style.

Table 3 Summary of -des- and -22245- in the polite style

 

 

 

 

 

 

 

 

 

 

 

 

 

copula (noun 8: 224 adjective)
affirmative negative
non-past -desu -j4 4rim4$e22
-j4 224i desu
past -de5bit4 -j4 4rim45e22 desbita
-j4 224/e4tt4 desu
polite ending (2' adjective)
non-past -i-desu Jeu arimasen
Jeu 224i desu
past -/e4tt4 desu Jeu arimasen desbita
Jeu 224/e4tt4 desu
verb
non-past 4224524 -m45e22
past -m45bit4 -m45e22 desbit4

 

 

 

 

In section 4.2., the effect of speech style on the use of fillers will be discussed.

Chapter 2
Review of previous studies

In this chapter, the deﬁnitions of the terms ‘discourse markers’ and ‘fillers’ are
first provided, since fillers are regarded as a type of discourse markers in this study.
Second, the properties of the discourse markers in English which are relevant to the
present study are identiﬁed. Third, major studies on Japanese ﬁllers are reviewed.
The last section is a summary.

2.1 . Definitions of ‘discourse markers’ and ‘fillers’

Schiffrin (1987) considers discourse markers to be important devices which
contribute to discourse coherence. As briefly mentioned earlier, Shiffrin gives an
Operational deﬁnition of discourse markers, ‘sequentially dependent elements which
bracket units of talk’ (p. 31), and speciﬁes the following conditions for a discourse
marker (p. 328):

1. Syntactically detachable from a sentence.

2. Commonly used in initial position of an utterance.

3. Has a range of prosodic contours, e.g. tonic stress and followed by a pause,
phonological reduction.

4. Operates at both local and global levels of discourse and on different planes
of discourse.

The above deﬁnition and conditions are adopted in this study as have other researchers,
such as Maynard (1989), Onodera (1993), and Philips (1998).
According to Maynard (1989:30), ‘the term “filler” refers to a broad range of

utterances (including independently appearing particles) that do not carry identifiable

10

or relevant propositional meanings. Some ﬁllers are phrases or brief comments, as
exempliﬁed by nan/e4 “well”; some are mere empty utterables such as uuuun “uhh”.’
Maynard’s deﬁnition of the term ‘ﬁller’ was employed in the present Study.

There are several classes of discourse markers other than ﬁllers, including
connectives, adverbs and interactional particles (also called final particles), among
others. Examples of connectives are sorede ‘and’ and demo ‘but’; an adverb, yappari ‘as
I thought’; and interactional particles, 22e ‘right?’ and yo ‘I’m telling you’.

The criteria researchers use to categorize words as ﬁllers varies. For example,
Philips (1998) considers certain interjections (e.g. [2024 ‘there!’) and adverbs (e.g. yappari
‘as I thought’) as fillers but this study focuses only on the ﬁllers specified in section 1.1.
2.2. Previous studies on discourse markers
2.2.1. Sacks, Schegloff and Jefferson (1974)

Sacks, Schegloff and Jefferson (1974) were among the first to explore discourse
markers in their research on turn taking. They called discourse markers ‘appositional
beginnings’ and state that ‘well, but, and, and so are extraordinarily common...
Appositional beginnings do not reveal much about the constructional features of the
sentence to follow’ (p. 32). In other words, they do not require that the speaker have
a plan in hand as a condition for starting. They further state that ‘appositionals then
are turn-entry devices—prestarts, just as the tag questions are exit devices—
postcompleters. Appositionals and tag questions are heavily used devices, though the
basis for their use is by no means self evident linguistically’ (p.32). As is apparent in

the next subsection, their study is elaborated by Owen (1981).

11

2.2.2. Owen (1981)

Owen gave additional consideration to Sacks, Schegloff and Jefferson’s (1974)
work on turn taking and conversational units. She (1981:108-112) analyzes the use of
well in terms of its distribution and functions. Limiting her work to instances of well
which preface the second pair-part of an adjacency pair, she found four types of second
pair parts: 1) Assessment-disagreement, 2) question-response, 3) request-noncompliance,
4) offer-rejection. She concludes that well is used to preface a face-threatening act, as a
strategy to signal that a face-threat is about to occur, reducing the subsequent threat.
She also emphasizes that the usage of well can only be analyzed in terms of the units of
real conversation.

As will be shown in section 4.4. Owen’s account of well that it is used to
reduce a face threat can be applied to the use of a certain type of Japanese ﬁllers,
especially when the psychological distance between the speakers is great. Her analysis
focuses on the pragmatic effects of an utterance, i.e. face-threat, while this study focuses
on the speakers’ psychological distance. The data shows that ﬁllers occur not only in
situations when a face-threat is about to occur, but also in positive situations such as
praising and congratulating. This point will be discussed further in section 4.4.

2.2.3. Schourup (1982)

While what has been predominantly discussed in the literature on discourse
markers is what is displayed or observable in an utterance, Schourup (1982) paid
attention to what is not displayed or observable in an utterance. He analyzed the

discourse functions of particles like, well and you know to identify a core use for each

12

one in relation to indications of what is unsaid. He proposes tripartite worlds in a
conversation (p. 7): ‘the private world’, referring to the covert thinking of the speaker;
‘the shard world’, referring to what is available to both the speaker and the bearer; and
‘the other world’ referring to the covert thinking of the other conversants which is
inaccessible to the speaker. He claims that unexpressed thinking of the speaker, ‘the
private world’, can be indicated by discourse particles (p. 18). He terms these
discourse particles ‘envincives’, deﬁned as follows:

Envincives: a linguistic item that indicates that at the moment at which it is

said, the speaker is engaged in, or has just then been engaged in, thinking; the

envincive item indicates that this thinking is now occurring or has just now
occurred but does completely specify its content.
He then states that ‘envincives are capable of two functions in conversation: they
establish the conversational relevance—but not the details—of undisclosed thinking by
the speaker, and they can mark the real time moment of occurrence of that thinking in
order to establish the timeliness of a speaker’s reaction’ (p. 21).

According to Schourup (p.64-66) well is an envincive with a particular basic
use, and this fact is crucial to understanding the various ways in which well is used in
conversation. He proposes that the core use of well, as the basic envincive use, is to
indicate that the present speaker is now examining the contents of the private world.
He examines the use of well in conversation by attempting to answer two questions: (1)
Where in the sequential development of a particular sentence, utterance, exchange,

conversation, etc. , does well occur? and (2) why does the speaker in that particular

context and sequential position choose to draw attention to his/ her examination of the

13

private world? He concludes that the primary use of well is an envincive indicating
consultation by the speaker of his/ her current thoughts (p. 91). A separate secondary
function depends on particular discourse environments.

Although Schourup’s approach may seem rather mentalistic, his unique
analysis of discourse particles as envincives is useful in this analysis of Japanese ﬁllers.
Shourup’s analysis relates what is covert to what is overt in ongoing conversation and
identify each particle has a single core use with a local function in conversation. This
is useful in the present study, as there is a relation between the use of fillers and
psychological distance between the speakers as represented by the choice of polite or
plain endings and such psychological distance maybe interpreted as ‘what is unsaid.’
2.2.4. Stubbs (1983)

Stubbs (1983:68-70), building on Sacks, et al (1974), analyzes discourse particles
such as well, now, right, OK, 422yw4y, you know, I see, hello, hyehye. He lists the
following characteristics of these particles:

1. They have no propositional meaning or truth value,

2. are essentially interactive and restricted to spoken language, and

3. when used utterance-initially, they relate utterances to each other or to
mark a boundary in the discourse.

He then examines well to illustrate his point. The functions Of well as an utterance-
initial particle include a general introductory function (Well, what shall we do?) and also
indicate an indirect answer when it is used after a question, claiming relevance although
admitting a shift in topic. He states that, although well and related items are not well

accounted for, one of their main functions is to act as boundary markers indicating the

14

boundaries of units larger than clauses or sentences.

Stubbs analyzes well from the perspective of discourse analysis, stating that
sentence grammars cannot sufﬁciently account for all the phenomena in discourse. In
other words, certain phenomena involving particles, or discourse markers and Others
can only be explained with reference to the unit at the clause level or above, which is
the position taken in the present study.

2.2.5. Schiffrin (1987)

Shiffrin (1987) is an influential work on discourse markers, integrating much
of the previous work. Since the outline of her work is presented in Section 1.3, this
subsection briefly reviews the functions of three discourse markers, well, y’know and I
mean, in order to illustrate her analysis.

According to Schiffrin (1987: 102-127), well is a response marker, which is
possible wherever the options offered through a prior utterance for the coherence differ
from an upcoming response. Thus, its main function is in the participation
framework of discourse. At the same time, well also operates in the information state
and the action structure since conversants can respond to anything in talk which
temporarily disrupts their expectations, for example, an unexpected utterance and a
request with no comply.

Y’know and I mean are characterized by Schiffrin as markers that operate on
both the information and the participation planes. While I mean focuses on the
speaker’s own adjustments in the production of his/ her talk, y’know proposes that a

hearer adjusts his/ her orientation toward the reception of another’s talk. Thus,

15

y’know works primarily in the information state and secondarily on the participation
framework, while the functions of I mean are reverse. Shiffrin’s method analyzing
these discourse markers will be applied to the analysis of Japanese fillers in section 4.4.
2.2.6. Holms (1986)

Holms (1986) examines, from a sociolinguistic perspective, the distribution
and functions of the expression you know in the speech of men and women. She
categorizes you know into two distinct functions which are then classified into a
number of subcategories (p. 7). In Category I, you know expresses speaker’s
conﬁdence or certainty. It comprises two subcategories: 1) you know expresses the
speaker’s certainty and conﬁdence about the addressee’s relevant background
knowledge and anticipated response, and 2) it positively reassures the addressee of the
validity of the proposition, where its function is emphatic. In Category II, you know
expresses uncertainty of various kinds. It also comprises two subcategories: 1) it
expresses the speaker’s uncertainty or lack of conﬁdence concerning the addressee’s
attitudes and likely response, and 2) it expresses the speaker’s uncertainty regarding the
linguistic encoding of the proposition.

The distribution and the functions of you know as analyzed above are similar
to those of ﬁllers like 4220 ‘well’, eeto ‘well’ and nanka ‘like’ in Japanese as will be
discussed in Chapter 4. Holm’s Categories I and II are useful in the present study,
since it can be assumed that when the speaker’s certainty toward the preposition is
higher, the psychological distance he/ she feels toward the addressee may be shorter,

and that when it is lower, the psychological distance may be greater. Holm’s notions

16

of uncertainty and certainty are related to Shiffrin’s ‘participation structure’, since they
deal with speaker-utterance and speaker-heater relationships.
2.2.7. Bloch (1996)

From a sociologist’s point of view, Bloch (1996) investigates the ways in which
emotions are expressed in narratives and conversations. She analyzes the narratives
and conversations for a systematic relationship between ‘paralinguistic markers’, on
one hand, and ‘positive and negative self-feelings’ on the other. In her analysis, ﬁllers
are regarded as markers of ‘negative self-feeling’, since they normally index uncertainty
and block the ﬂow of speech. Bloch’s notion of ‘negative self-feeling’ is relevant to
psychological distance in this current study. In other words, when the psychological
distance between the speakers is greater, the speaker’s negative self-feeling may be
greater; however, Bloch’s contention that ﬁllers ‘normally block the ﬂow of speech’ is
too simplistic, because Japanese ﬁllers commonly seem to help, not block, the flow of
speech. These points will be further discussed in Chapter 4.

2.3. Previous studies on fillers in Japanese
2.3.1. Maynard (1989)

Maynard (1989) presented a pioneering study in the ﬁeld of Japanese
conversation analysis. According to her (p.30), ‘the frequent use of fillers marks a
noticeable feature of conversational discourse.’ She categorizes ﬁllers into two groups:
(1) language-production-based ﬁllers, and (2) socially motivated ﬁllers. The ﬁrst type
appears when smooth speech is either cognitively or productively hindered. They are

typically accompanied by meta-communicational comments, such as uunto, are, hora

17

‘Uhh...that’. She further states that ‘fillers also create an atmosphere in which the
speaker shows some hesitancy and less certainty about the message content and,
therefore, gives an impression of being less imposing’ (p. 31). She points out that
fillers, as well as other devices, play an important role in achieving maximum
agreeableness to the recipient, ensuring interpersonal rapport. She also states that the
frequent use of ﬁnal particles and ﬁllers makes certain that the interpersonal feelings are
intact when the semantic content is conveyed to the other interactant (p.31).

While Maynard’s (1989) observations on the interactional effects of fillers seem
valid, she does not use naturally-occurring data nor does she identify what should be
considered as fillers, the environments in which they occur, or their frequency. Thus,
there is a need for an empirical study to understand the facts about ﬁllers in Japanese,
which is the main purpose Of the present study.

2.3.2. Saito (1992)

Saito (1992) analyzes nanka ‘like’ as a discourse marker in Japanese
conversation. According to her, nanka has several functions: (1) utterance softener,
(2) a turn initiator, and (3) a ﬁller [language-production—based ﬁller] (p.58). Although
she claims that the occurrence of nanka may be sensitive to the degree of how relaxed
the speaker feels, among other factors, she does not elaborate on it, much less present
quantitative data. In section 4.2, the number of ﬁller occurrences will be compared
quantitatively with those of polite expressions, which reﬂect the psychological distance
the speaker feels toward the addressee. Japanese fillers of both language-production

and socially motivated types, including nanka ‘well’, based on the data will also be

18

analyzed.
2.3.3. Cook (1993)

Cook (1993) analyzes the use of 4220 ‘well’ as a ﬁller in terms of how it indexes
politeness. Five uses of the ﬁller 4220 based on her data (one family conversation and
one Diet interpellation) are listed as follows: (1) conversation starter or turn-initiator,
(2) attention getter, (3) proposition highlighter, (4) new topic starter, and (5)
disagreement softener (p. 23). Cook concludes that the function of 4220 is to obtain
the addressee’s cooperation by evoking the feeling that the interlocutors are on the
same side, which comes from its deictic meaning as demonstrative adjective, with
respect to the subsequent utterance. In this way, 4220 is used as a politeness marker
which aligns both the speaker and the addressee (pp. 22-31).

Although Cook (1993) uncovered interesting facts regarding 4220, its functions
need to be further examined. For example, she interprets 4220 as a positive politeness
marker, which seems to suggest that she regards it only as what Maynard (1989) calls a
‘socially motivated filler’. In the present analysis, however, certain instances of 4220
can actually be regarded as what Maynard calls a ‘language-production-based ﬁller’. A
more detailed discussion on this matter will be provided in Chapter 4.

2.3.4. Philips (1998)

Philips (1998) presents a comprehensive analysis of discourse markers in
Japanese, classifying them into three types including connectives, fillers and
interactional particles. Fillers are analyzed in Philips’ study in terms of the degree of

formality, as well as age, gender and speech genre. She ﬁnds that the rates of ﬁller are

19

much higher in formal speech than in informal speech, and that types of ﬁllers were
also inﬂuenced by the degree of formality (p. 113). In her study, speech genre
(interviews, conversations, and narratives) does not inﬂuence the use of the ﬁllers (p.
119). She concludes that Japanese fillers help to create a socially and culturally
expected interactional atmosphere by showing hesitation and uncertainty (p.261).

In the present study, ‘formality’ is interpreted in terms of the psychological
distance between speakers, i.e. formal style is used in conversation when the
psychological distance is great. Other means of measuring psychological distance in
addition to the ones in Philips’ study are employed. In other words, the answers on
the questionnaire which the subjects ﬁlled out regarding the closeness toward the
interlocutor and their impression of the conversation and of the interlocutor, in order
to measure psychological distance are employed in the present study.

2.4. Summary

In this chapter deﬁnitions of discourse markers and ﬁllers were provided.
Previous studies on ﬁllers and discourse markers in both English and Japanese were
also reviewed. A summary of previous studies on discourse markers in English is as
follows: Sacks, et a1 (1974) analyzes discourse markers as ‘turn-entry devices’; Owen
(1981) analyzes well as softening device for a face-threatening act; Shourup (1982)
analyzes discourse markers as ‘envincives’; Stubbs (1983) analyzes well as a boundary
marker; and Shiffrin (1987) analyzes discourse markers as ‘contextual coordinates’
which contribute to the discourse coherence. Two studies on discourse markers in

English from sociolinguistic and sociological perspectives were also reviewed.

20

The previous studies on Japanese ﬁllers are summarized as follows: according
to Maynard, Japanese ﬁllers are devices to achieve maximum agreeableness to the
recipient; Saito (1992) observes that nanka ‘kind of’ occurs when the speaker feels
relaxed; Cook (1993) analyzes 4220 ’well’ as a politeness marker; and Philips (1998)
analyzes Japanese ﬁllers as markers which contribute interactional atmosphere.

While Saito’s (1992) and Philips’ (1998) studies on Japanese ﬁllers treat ﬁllers
in relation to the speaker’s psychological state, no study has been conducted on a
possible relation between psychological distance between speakers and the use of ﬁllers.

It is hoped that the present study will be useful in understanding Japanese fillers.

21

Chapter 3
Methods and procedures

3.1. Data

The procedures for data collection in the present study are based on those by
Maynard (1989:15-18). To examine the actual language use, naturally-occurring
conversations are used as the data. Twelve dyadic conversations between native
speakers of Japanese were audiotaped for the present study. The participants were
aware that they were being audiotaped, but the particular focus of the present study
was not revealed. Some participants are close friends with their interlocutors while
others are just acquaintances or total strangers. Pairs varying in the degree of
closeness were included in order to measure psychological distance. The participants
were asked to talk about two topics including their childhood dreams and their views
on the differences in education in the US. and Japan. The participants were free to
choose additional topics. Most of the conversations took place in the participants’
houses and at public places, such as coffee shops or their places of work. The
participants were asked to refrain from eating or drinking during the taped
conversation since it might affect the occurrences of ﬁllers. The researcher was absent
during the conversations, each of which lasted about 15-30 minutes. At the end of
each conversation the participants completed a questionnaire regarding'their
backgrounds, the relationship with the interlocutor and their impressions of the
interlocutor and of the conversation that had just taken place. The latter two are

intended to measure the psychological distance toward the interlocutor and the

22

participants’ comfort level during the conversation (see Appendix B for the
questionnaire).
3.2. Participants

Twenty-one women participated in this study, all of whom were native
Japanese speakers. The participant were paired up with the conversation partner
whom the researcher assigned. Three participants participated in two conversations.
The present study does not concern gender differences in the use of ﬁllers. The
occupations of the participants varied. Some were students, some were housewives,
and others were teachers. They ranged in age from their early twenties to early forties.

Out of twelve, one conversation was recorded in Japan while the rest were recorded in

the US. Table 4 summarizes the participants’ proﬁles.

Table 4 Participants

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

identiﬁcation age occupation identification age occupation
1-YM 41 student 12-TH 37 housewife
2-IS 35 housewife 13-MR 27 teacher
3-OT 26 student 14-IY1 32 teacher
4-MSl 22 student 15-FY 31 teacher
5—ON 31 student 16-IM 31 teacher
6-JY 39 housewife 17-IT 30 student
7-M82 35 student 18-AY 29 student
8-OM 37 student 19-IY2 25 student
9-NH 37 housewife 20-OY 23 student
10-TM 41 housewife 21-88 28 student

1 l-TS 37 housewife

 

 

 

 

3.3. Procedures

After collecting the data, the first ten-minute segment of each conversation was

23

 

fully transcribed. Although the first two-to-three-minute segment is commonly
excluded by other conversation analysts, it is purposely included in the present study in
order to verify discomfort of conversing between pairs of strangers or acquaintances
due to psychological distance and which was reflected in the occurrences of ﬁllers.
For transcription, Maynard’s (1989) conventions were adopted (see Appendix A). In
total, 120 minutes of the conversations were transcribed and analyzed.
3.4. Methods of analysis

The number of occurrences of each ﬁller was totaled and the frequency of each
ﬁller per 1000 words (‘the rate’ hereafter) was calculated by dividing the number of
occurrences of each filler by the number of words uttered, and then multiplying the
result by 1000.

The data was analyzed in terms of the following characteristics:
(1) The total number and the rate of the occurrence of each ﬁller;
(2) the relation between the frequency of ﬁllers and the speech styles (polite vs. plain);
(3) the relation between the frequency of fillers and psychological distance between

interlocutors as reﬂected in the questionnaire results;
(4) the environments in which fillers occur; and

(5) the function(s) of ﬁllers.

24

Chapter 4
Results and analysis
4.1. Total number and the rate
Among all ﬁllers, nanka ‘like’ was used most frequently, followed by 4n0(0)
‘well’. Of 608 tokens, 252 (41.4%) were nanka ‘like’ and 213 (35.0%) were 4n0(0)
‘well’. These two constituted over 75% of all fillers in the data. The least frequently

used ﬁller was k0n0 ‘well’. Table 5 summarizes the number and the rate of fillers.

Table 5 Summary of ﬁller

 

 

 

 

filler meaning number of tokens percentage’ rate/1000 words 2

nanka like 252(16) 41.4% 12.1
4n0(0) well 213(6) 35.0% 10.2
son0(0) well 35(6) 5.8% 1.7
maa kind of 33(1) 5.4% 1.6
ee(e) uh 26(3) 4.2% 1.2
leetto’ well 19(0) 3 . 1% 0.9
uun(t0) well 16(3) 2.6% 0.8
iy4(4) well 14(3) 2.3% 0.7
44 uh 0(13) 0% 0
k0n0(0) well 0(1) 0% 0
total 608(52) 100% 29.2

 

 

 

 

 

 

 

As mentioned in section 1.4., there are quite a few occurrences of words (52 tokens) in
the data that cannot be clearly identified as ﬁllers. These ‘fuzzy’ words appear in the
parentheses in the ‘number of tokens’ column of Table 5. In order to classify them

into ﬁllers or other categories with conﬁdence, it would have been necessary to obtain

 

1 do not include fillers in () 2 do not include fillers in () 3 includes variants eta, eeto, and etto

25

additional information from the participants about what they meant at the time of

recording, which was impossible. Moreover, a close examination of the context (time,

place and what was previously discussed, among others), the turns, the accent and

intonation of the word would have also been necessary. Therefore, these words are

left uncategorized in the present study.

Example (2) illustrates the occurrence of 4220 ‘well’ that cannot be categorized

as either a ﬁller or a demonstrative adjective in its anaphoric use.

(2) T:

a. Demo ano juken to wa zenzen chigau ja II.
but F? entrance exam with Top totally different TAG

b. juken no jukenbenkyoo ja nai ja n zenzen
entrance exam GEN study for exam COP NEG TAG totally

c. ano tsumekomu n ja naku te motto minna de...
F? cram NOM COP NEG and rather everyone with

d. ...ano juken ne ano gaa tte tsumekomu tte yuu
F? entrance exam IP F? QT cram QT

e. arm a yuu noo no tsukai kata shi nai yo ne.
F? that way brain GEN use way do NEG IP IP

‘But, it’s totally different from taking an entrance exam in Japan, isn’t it? It’s
not like studying for an entrance exam. It’s not cramming, rather, it’s like
(thinking) together... Students here use their brain differently from students
in Japan’

In (2), all the occurrences of 4220 can be interpreted as fillers, ﬁlling pauses, or as

demonstrative adjectives, pointing out something in the context (in this particular

example, pointing out something in common in interlocutors’ background

knowledge).

26

4.2. Fillers and speech styles
Maynard (1989:30) categorizes ﬁllers into two groups based on the motivation
for their use: (1) language-production-based ﬁllers, and (2) socially motivated ﬁllers.

She states:

the ﬁrst kind appears when smooth speech is either cognitively or
productively hindered. They are typically accompanied by meta-
communicational comments, such as uunto, are, hora ‘Uhh...that’... As
frequently used is the second category...These ﬁllers may be employed to ﬁll a
potential silence. Uttering mere sounds makes it possible to create the
impression that verbal interaction is carried on without cessation, thereby
avoiding potential embarrassment.
Both language-production-based ﬁllers and socially motivated ﬁllers were observed in
the current data, although some ﬁllers were difficult to classify into one category or the
other‘ . The criteria used to categorize ﬁllers as one of the language-production-based
types is as follows: (1)when a filler appears with stammering or faltering, including
when the speaker corrects him/ herself (‘self repair’ hereafter); and (2) when a ﬁller
appears with other phrases that means the speaker is trying to recall the words, nan d4
kke ‘what was it?’ or nan te yuu n da kke ‘what do you call it?, etc. All other ﬁllers
not categorized as the language-production-based type are categorized as the socially
motivated type. Although all fillers can potentially function both as language-
production-based types and socially motivated types depending on the context, in the

present data, only nanka ‘kind of’, an0(0) ‘well’, and eeto ‘well’ were used as both types.

Table 6 summarizes the kinds of ﬁllers for both types.

 

’ Categorization of fillers can be further studied.

27

Table 6 Language-production-based type 86 socially motivated type

 

 

 

Icategories ﬁllers

Ilanguage-production—based a220, nanka, eto

socially motivated an0(0), son0(0), nanka, iya(a), maa, ee(e), eetto’ ,
uun(t0)

 

 

 

The examples below illustrate the occurrence of the language-production-based ﬁller
4220 ‘well’, shown in bold. (3) and (4) are the examples of the criteria (1) and (2),
respectively.

(3) T: a. Nde nanka ma ano shi— ano ronbun.
and F PF PF PF thesis

‘And like, well, uh, the thesis.’

(4) M: a. Ano saa ano ba- ano nani butai ka nanka no bakku de
F IP PF PF what Stage Q something GEN back at

b. odoru dansaa tte yuu ka saa.
dance dancer QT say Q IP

‘Well, uh, uh, it’s like the dancers who dance behind the singer on the stage.’
A detailed discussion on the environment in which language-production-based fillers
occur will be presented in section 4.4.

Example (5) illustrates an occurrence of a socially motivated filler, shown in

bold.

(5) T: a. Anoo moo daigaku wa de rarete daigakuin
SF already college TOP graduate HON graduate school

b. ni sugu desu ka
to right after COP Q

‘Uh, did you already graduate from college and go to graduate school right
after that?’

 

5 includes variants eta, etto, eeto

28

There are occurrences of ﬁllers that cannot be categorized into either the
language-production-based type or the socially motivated type, just like the words that
cannot be categorized as ﬁllers or as words in another category. It is impossible to
categorize them with conﬁdence unless the researcher asks the participants about what
was happening in the speakers mind right at the moment when the ﬁller was uttered,
which was impossible to do. Thus, it is apparent that categorizing fillers into one of
the two types is not simple and clear-cut.

Of the 608 tokens in the current data, 23 belong to the language-production-
based type, 6 to the ‘fuzzy’ type, and the remaining, 579, to the socially motivated type.
Table 7 summarizes the occurrences of all types of fillers.

Table 7 Number of occurrences of each type of ﬁller

 

 

 

 

 

 

We of ﬁller tokens
socially motivated ﬁller 579(95.2%)
language-production-based ﬁller 23(3.8%)
“fuzzy” ﬁller 6(1.0%)
total 608(1000/o)

 

 

 

Example (6) illustrates the ‘fuzzy’ filler that cannot be categorized as either
types, shown in hold.

(6) S: a. Aa fagotto a $00 ka.
oh bassoon oh so Q

b. Are dooyatte oto dasu n desu ka.
that how sound make NOM COP Q

M: c. Are anoo are wa ooboe to koko wa kuchi n toko
that PF? that TOP oboe as here TOP mouth GEN part

(1. wa onnaji de tada ookii dake da to omou n desu
TOP same and just big only COP QT think NOM COP

e. kedo.

29

but

‘Oh, bassoon, that’s what it is. How do you make a sound with it?’
‘Its mouth is same as oboe’s. It’s just bigger than that, I think.’

In (6) above, the speaker repeated the word are ‘that’ (in line c) and the topic marker
wa was repeated three times with different phrases (in line c-d), which shows hindering.
However, no false start or self repair was observed. In this case, the filler 4220 ‘well’
can be interpreted two ways. The speaker was either physically searching for a word
or a phrase in what follows (language-production-based), or the speaker was searching
for the appropriate words or phrases in the following part of the sentence not to be
imposing to the addressee with a social motivation. It seems that the environment to
judge this filler as either a language-production-based type or as a socially motivated
type is not strong enough.

The rates of ﬁllers in each speech style are now examined. It is reasonable to
assume that the more nervous or uncomfortable the speaker is, the more language-
production-based fillers occur, since discomfort hampers smooth speech. Thus, to
analyze a relation between psychological distance between the speakers and the use of
fillers, it was decided that the number of occurrences of both language-production-
based fillers and socially motivated ﬁllers should be used.

Now the relation between the occurrence of ﬁllers and the speech style used is
examined. The data from the present study showed that the subjects conversing with
a person who they met for the ﬁrst time talked predominantly in the polite style, as
expected. Example (7) illustrates the occurrence of the polite form, shown in bold.

(7) S: a. De okosan wa moo ookikute irassharu n desu ka.
and child TOP already big HON NOM COP Q

30

M: b. Kodomo wa ne ano nanasai desu.
child TOP IP F seven years old COP

‘Is your child already old enough?’
‘She is, well, seven years old.’

In 4 out of 12 conversations the polite style was used predominantly, and in all
of these cases the interlocutors were strangers, meeting each other for the first time.
The rate of occurrence of the polite form was calculated by dividing the number of
occurrences of -des- and -mas- by the number of sentence ﬁnal position which -des- and -
mas can possibly occur. Table 8 summarizes the occurrences of the polite forms.

Table 8 Occurrences of -des- and -mas- in each conversation
°onshi nversation ID . of -des-&-mas- rate lite form

7 150 .8°/o

8 130 .9°/o

12 39 .5%
3

\l
0

0.2%
0. 1%
0. 1%
0%
0%
0%
0%
0%

friends
’ 10
friends 4
2
ose friends 5
' tances 6
9
ose friends 14

 

3
2
2
0
0
0
0
0

As can be seen in Table 8, the pairs in which interlocutors were strangers used polite
forms predominantly. These four conversations were classiﬁed as the ‘polite type’
and the remaining conversations as the ‘plain type’. Table 9 compares the occurrences

of ﬁllers in each speech style.

 

" -des- and -mas- in quotation is not included

31

Table 9 Occurrences of fillers in each speech style

 

 

 

 

I polite type plain type
[total number of words occurred 7119 13683
Inumber of ﬁllers 263 345
[Late of ﬁller / 1000 words 36.9 25.2

 

 

 

 

Table 9 shows that the rate of ﬁllers is higher in the polite style conversation (37/ 1000
words) than in the plain type (25/1000 words). Thus, it is concluded that the speech
styles do affect the occurrence of ﬁllers. In other words, more fillers occur in the
polite style conversation, in which psychological distance towards the addressee is
assumed greater than one in the plain style. I speculate this is because the speaker
must think more carefully about what to say and how to say it so that ﬁllers occur
frequently thus sounding less imposing and avoiding potential embarrassment.

The kinds of ﬁllers that were used also varied in the two speech styles. The
casual sounding ﬁller nanka ‘like’ occurred more frequently in the plain type.
Relatively formal sounding fillers, such as 4220 ‘well’ and sono ‘well’, on the other hand,
occurred more frequently in the polite type. A similar result was obtained by Saito
(1992). In her study, nanka ‘like’ occurred more frequently in same gender group
which speakers were more relaxed than mixed gender group. Table 10 summarizes

the kinds of ﬁllers used in the polite and plain styles.

32

Table 10 Summary of kinds of fillers for each speech style

 

 

 

 

 

 

I polite type plain type

total no. of words 7119 13683

ﬁller meaning tokens rate/1000wrds tokens rate/1000wrds
4220(0) well 118 16.4 95 6.9
son0(0) well 18 2.5 17 1.2
k0n0(0) well 0 0 0 0
nanka like 57 8.0 195 14.3
iya well 6 0.8 8 0.6
44 uh 0 O 0 O
maa kind of 16 2.1 17 1.2
ee(e) uh 25 3.4 1 0.1
eetto’ well 16 2.2 3 0.2
uun(t0) well 7 1.0 9 0.7
[total 263 36.9 345 25.2

 

 

 

 

 

 

 

Example (8) and (9) illustrate the use of ﬁllers, shown in bold, in each speech style. (8)
and (9) are the examples of polite and plain style, respectively.

(8) Y: a. Anoo kookoo no toki no chiri no jikan ni
F high school GEN time GEN geography GEN class in

b. ano suraido suraido no Shoo ga atte sore ga
F slide slide GEN Show SUB have that SUB

c. nyuujiirando datta n desu.

New Zealand COP NOM COP

‘Well, in geography class when I was in high school, well, they had a slide
show and it was about New Zealand.’

(9) S: a. Nichibee no kyooiku no chigai nanka
Japan and US GEN education GEN difference F

b. konaida batorukuriiku no hoshuukoo de nee
the other day Battle Creek GEN Japanese School at IP

c. kakaseta no yo nanka kookoosee ni.

 

7 includes variants eta, etto, eeto

33

write-had NOM IP F high school student P

‘Well, the other day, well, I had high school students write about the
differences in education between Japan and US. at Japanese School of Battle

Creek.’
4.3. Fillers and questionnaire results

In this subsection, the relation between the use of fillers and questionnaire
results is examined. Immediately after recording the conversations, participants ﬁlled
out a questionnaire regarding their backgrounds, the relationship with the interlocutor,
and the impressions of the conversation and of the interlocutor (see Appendix B).
Answers were coded as numbers, and the relationship between interlocutors, the
participants’ comfort level during the conversation, and the psychological distance
toward the interlocutor were examined.
4.3.1. Fillers and relationship between interlocutors

Two participants were paired up with their best friends, two were with close
friends, two were with just friends, two were with acquaintances and four were persons
who they met for the first time. Fillers occurred most frequently in the conversation
between strangers, which conﬁrmed the hypothesis that the more the speaker feels
psychologically distant from the addressee, the more frequently fillers are used.
Fillers occurred least frequently in the conversation between acquaintances, however,
contrary to the expectation that it would be between best friends. In general, a clear
relation between the frequency of ﬁllers and the closeness between the interlocutors
was not observed. This may be due to the fact that factors other than the closeness

between the interlocutors had a stronger affect on the occurrence of fillers. For

34

example, the pair in conversation 8, even though the interlocutors met for the first
time, seemed to enjoy the conversation very much because they found many things in
common in their background. Therefore the answers by both speakers regarding the
comfort level were high. Thus, it is not accurate to automatically assume that the
relative closeness of people who meet for the ﬁrst time is always low. Furthermore,
the interlocutors’ interpretations of the terms ‘acquaintance’, ‘friend’, ‘best friend’ and
so forth, which they were asked to use to describe their relationship with the
interlocutor in the questionnaire, may have varied among the participants. Table 11
summarizes the relation between the number and the rate of ﬁllers used and the
closeness between interlocutors, as described by the participants.

Table 11 Fillers and relationshi 3 between interlocutors

 

 

 

 

 

 

 

 

 

 

 

 

 

conversation ID relationship no. of ﬁller rate/1000 words
7 strangers 78 43.8
13 best friends 56 41.9
3 strangers 80 41.4
14 close friends 64 39.9
10 acquaintances 47 34.8
8 strangers 69 32.9
12 strangers 36 27.48
5 close friends 50 27.46
4 best friends 45 21.1
9 friends 27 16.0
6 acquaintances 3 1 15 .7
2 friends 22 12.9

 

 

 

 

 

 

4.3.2. Fillers and impressions of conversation and of interlocutor
Answers from the participants regarding the impression of the conversation
and of the interlocutor were coded as numbers. The numbers in the answers on the

questionnaire were added to measure the impressions (12 - most relaxed and 60 - most

35

nervous). The fillers occurred most frequently in the conversation of the pair who
felt relatively relaxed and regarded the conversation pleasant, rather than in the
conversations of the pair who were most nervous and regarded the conversation as
unpleasant. This was contrary to the expectation. The ﬁllers occurred least
frequently in the conversation of the pair whose comfort level was in the middle range
among all the participants. Again, a clear relation between the use of fillers and the
impression of the conversation and of the interlocutor was not observed. This may be
due to the fact that factors other than the impressions of the conversation and of the
interlocutor affected the occurrence of fillers. For example, one participant who used
nanka ‘like’ most frequently did so no matter what the topic was, talking to her best
friend. This participant had a habit of saying nanka ‘like’. It is also possible that the
interlocutors’ impression may not have been accurately reﬂected on the answers in the
questionnaire even though ‘this information will be kept strictly conﬁdential and
please answer honestly’ is marked on the questionnaire. First of all, if a participant
felt nervous, she may not have revealed it on the questionnaire, and even if she had
expressed it, the degree chosen may not have been accurately reﬂected on the
questionnaire. Secondly, interpretations of the terms ‘the conversation went
smoothly’, ‘I was relaxed’ and ‘the conversation was pleasant’ and so forth, which the
participants described their impressions of the conversation and of the interlocutor in
the questionnaire, may have varied among the participants. Table 12 summarizes the
relation between ﬁllers and the impressions of the conversation and of the interlocutor

in order of decreasing degree of discomfort.

36

Table 12 Fillers and impressions of conversation and of interlocutor

 

 

 

 

 

 

 

 

 

 

 

 

 

lconversation ID impression no. of filler rate/1000 words
6 43 31 15.7
3 24 80 41.4
14 21 64 39.9
10 20 47 34.8
9 20 27 16.0
2 18 22 12.9
5 17 50 27.46
12 16 36 27.48
7 15 78 43.8
8 13 69 32.9
4 13 45 21.1
13 12 59 41.9

 

 

 

 

 

 

The results show that the hypothesis indicating there is a relation between
psychological distance and the use of ﬁllers, more speciﬁcally, that the more the
speaker feels psychologically distant from the addressee, the more frequently ﬁllers are
used is partially supported. Although a relation between the frequency
of fillers and speech styles was observed, (i.e. in the polite style, more ﬁllers occurred),
between the frequency of ﬁllers and the relationship between the interlocutors was not.
Nor was a relation between the frequency of ﬁllers and the impressions of conversation
and of interlocutor.

4.4. Environments and functions of fillers

In this subsection, the environments in which ﬁllers of both language-
production-based type and socially motivated type occur are ﬁrst examined. Second,
the functions of ﬁllers are examined.

4.4.1. Environments of language-production-based fillers

Of the 608 tokens, 23 were language-production-based fillers (See Section 4.2.).

37

A220 ‘well’, nanka ‘like’, and eeto ‘well’ were used as language-production-based ﬁllers.
All of them were turn-medial, not turn-initial or ﬁnal. Categories were divided into
self repair and groping for a word, and each category was divided further into several
subcategories. Language-production-based ﬁllers occurred with self repair (17 tokens
74%) most frequently, and second most frequently when the speaker was groping for a
word (6 tokens 26%). Within the category of self repair, ﬁllers occurred with false
starts most frequently (12 tokens 52%), followed by the occurrences with wrong
choices of words (3 tokens 13%). Some fullers occurred with the word nani ‘what’
when the speaker was groping for a word (6 tokens 26%). Table 13 summarizes the
environments in which language-production-based ﬁllers occurred.

Table 13 Summary of environments of language-production-based fillers

 

 

 

 

 

 

 

 

 

 

 

category category subcategory tokens
self repair with false starts before false starts 7 (30%)
after false starts 5 (22%)
after ja nai ‘not’ 2 (9%)
with wrong choices of before wrong choices of 2 ( 9%)
words words
after wrong choices 1 (4%)
groping for a word with nani ‘what’ before nani ‘what’ 6 (26%)
after nani ’what’ 0
| total 23 (100%)

 

 

 

 

 

Example (10) illustrates the occurrence of a ﬁller (shown in bold) before a false start.

(10) S: a. Kisha ni naritakatta. Nanka tetsuga- koo porishii tte yuu
writer P become-wanted to PF this policy QT say
b. ka sorenari no tetsugaku ppoi no ga

Q in its own way GEN philosophy like NOM SUB

c. kaitearu janai desu ka.

written TAG COP Q

‘I wanted to become a magazine reporter. Well, you can tell their philoso-

38

like policy or philosophy is written (in the article), can’t you?’
Example (11) illustrates the occurrence of a ﬁller (shown in bold) after a false start.

(11) Y: a. De sonnani shi- ano shinkenni kangaete nai kara
and so much PF seriously think NEG because

b. kodomo no koro no yume.
child GEN time GEN dream

‘And because I didn’t take my childhood dream se-, well, seriously.’
Example (12) illustrates the occurrence of a ﬁller (shown in bold) before nanz' ‘what’.

(12) T: a. Demo mishigan de yuki suki toka anoo nani koo
but Michigan in snow like QT PF what this

b. donyorishiteru sora ga suki toka yuu hito imasu kara
gloomy sky SUB like QT say person there is because

‘But there are people who like snow and well, what, a gloomy sky.’
Example (13) illustrates the occurrence of fillers (shown in bold) before wrong choice

of words.

(13) Y: a. nanka oto nanka gakufu gakuten gakufu yomu no.

PF note PF score musical grammar score read NOM

111,8 like 1' eadin musical notes WCll, musical score, musical 1' ammar , musical
g 2
2
score.

Example (14) illustrates the occurrence of a filler (shown in bold) with ja nai ‘not’.

(14) S: a. aa aa aa nanka kanatteru ja n batoru ja nakute ano
ah ah ah F come true TAG Battle COP NEG PF

b. ringo de.
Ringo at

‘Your dream came true at Battle, not, uh, at Ringo School.’

39

4.4.2. Functions of language-production-based fillers

In the data, language-production-based fillers 4220 ‘well’, nanka ‘like’, and eeto
‘well’ primarily occur in the following contexts: (1) before and after self repair (14
tokens 61%); (2) while groping for a word (9 tokens 39%). The main function of a
language-production-based filler is a ﬂoor-holding device, signaling that (1) the speaker
is correcting him/ herself and/ or that (2) the speaker is searching for a word and
signaling his/ her intention to continue.
4.4.3. Environments of socially motivated fillers

The environments of socially motivated ﬁllers are now reviewed. Of 608
tokens, 579 were socially motivated ﬁllers (see Section 4.2.). An0(0) ‘well’, son0(0)
‘well’, nanka ‘like’, iy4(4) ‘well’, 22244 ‘kind of’ ee(e) ‘uh’ eetto ‘well’ and uun(t0) ‘well’
were used as socially motivated ﬁllers. In what follows, the environments of 4220(0)
‘well’ and nanka ‘like’ are examined since these two constitute over 7 5% of all ﬁller
usage in the data. Nanka ‘like’ occurred turn-initially (12.1%), medially (82.5%) and
ﬁnally (5.3%), while 4n0(o) ‘well’ occurred only turn initially (9.2%) and medially
(90.8%). Table 14 summarizes the occurrences of socially motivated fillers nanka
‘well’ and 4n0(0) ‘well’ with respect to the position.

Table 14 Summary of nanka ‘like’ and 4220 ‘well’ with position

 

 

 

 

 

I turn initial turn medial turn final total J
Inanka ‘like’ 30 (12.1%) 203 (82.5%) 13 (5.3%) 246 (100%)|
Ian0(0) ‘well’ 17 (9.2%) 168 (90.8%) 0 185 (100%)|

 

 

 

4.4.4. Functions of the filler nanka ‘like’

The functions of nanka ‘like’ are ﬁrst examined because it was the most

40

frequently used ﬁller in the data. Nanka ‘like’ is originally an indeﬁnite noun
meaning ‘something’ and constituting of nani ‘what’ and the question particle ka.
Even when it is used as a ﬁller, the original meaning plays a role of showing
uncertainty. Saito (1992:57) analyzes the functions of nanka ‘like’ as an utterance-
softener, a turn initiator, and a ﬁller (equivalent to language-production-based filler in
this Study). Philips (1998:136) analyzes nanka ‘like’ as a place holder, prefaces a
change in an utterance, to get the addressee’s attention, and a marker of hesitation.

Nanka ‘like’ in the present study primarily occurred in the following
contexts: (1) prefacing the main point of utterance; (2) prefacing the contents of an
utterance; and (3) initiating a turn. The contexts (2) and (3) conﬁrm Saito’s
observation. The contexts (1) and (3) conﬁrm Philips’. The main function of nanka
‘like’ thus seems to soften the content, showing hesitation and/ or uncertainty, and
trying to sound nonimposing to the addressee. Examples below illustrate these three
contexts.

First of all, nanka ‘like’ is often used to preface the main point of an utterance,
41 tokens (17%), as shown in example (15) in bold.

(15) M: a. Jissai wa sore sore bakkari ja nai n da yo
reality TOP that that only COP NEG NOM COP IP

b. tteyuu no 0 nanka jissaitekini oshienaito saa sore ga
QT NOM ACC SF realistically teach-must 1P that SUB

c. gakkoo jan.
school TAG

‘They have to teach (in America) that, well, in reality. You can’t be just that
(happy-go-lucky). That’s school’s responsibility, right?’

41

In (15), the main point of the utterance is that schools in the US. must teach reality.
The nanka ‘like’ in line b functions as a content softener, toning down a content of
strong criticism.

Secondly, nanka ‘like’ is often used to preface the contents of an utterance
other than the main point, 34 tokens (14%), as shown in example (16) in bold.

(16) S: a. Aa soo da ne shoogakkoo tsutte mo ne imaichi.
yes do COP IP elementary school QT P IP not sure

b. Nan datta kana watashi wa nanka yoochien no nanka
what COP Q I TOP SF kindergarten GEN SF

c. owari no koro ni shoorai nani ni naritai desu ka
end GEN time at future what P become want to COP Q

d. toka tte e kaite kudasai nante iwarete oboeteru n
QT picture write please QT say-PASS remember NOM

e. sore ni nanka hanaya ni naritai toka itte nanka...
besides SF ﬂower shop P become want to like say SF

‘Yes, it’s hard to remember what my childhood dream was when I was in

elementary school. Well, like, at the end of my kindergarten days, they

asked me what I wanted to become in the future, and asked me to draw a

picture about it. I remember saying that I wanted to be, like, a ﬂorist, like...’
In (16), the main point of the utterance is that the speaker desired to become a ﬂorist, as
indicated in line e. The occurrences of nanka ‘like’ other than the one prefacing the
main point, (in lines b and e, shown in bold) either Show the speaker’s hesitancy to
discuss her own personal past experience or are used as content softeners in order to
appear uncertain and not seem imposing.

Finally, nanka ‘like’ is often used at the beginning of a new speaker’s turn, 30

tokens (12%), shown in example (17), in bold.

(17) S: a. Iya demo koo yatte nanka mitsukete iku n ja nai.

42

no but this do something ﬁnd go NOM COP

T: b.Nanka yabai yo.
SF trouble IP

‘I think you will ﬁnd something along the way.’
‘Like, I’m in trouble.’

In (17), after S’s turn, T begins her turn with nanka ‘like’. Nanka ‘like’ occurs in
turn-initial position not to sound abrupt when initiating a turn. In addition, nanka
‘like’ in line b shows the speaker’s hesitancy to discuss her own personal trouble and
tone down the content of the strong word yabai ‘in trouble’.

Nanka ‘like’ is often accompanied by other softening devices (94 tokens 38%),
such as toka ‘kind of’, mitai ‘like’, tari ‘such as’, kanji ‘feel like’, and interactional
particle ne ‘right?’. This also supports my contention that the main function of nanka
‘like’ is as a content softener, showing hesitation and/ or uncertainty not to be
imposing.

In this subsection three contexts were examined in which nanka ‘like’ was
used: (1) prefacing the main point of an utterance; (2) prefacing the contents of an
utterance (other than main point); (3) initiating a turn. The main function of nanka
‘like’ is a softener, showing hesitation and/ or uncertainty not to be imposing to the
addressee. The data in this study conﬁrms Philips’ (1998) observation that nanka
‘like’ is a hesitation marker enlisting the addressee’s cooperation.

The following discussion examines how nanka ‘like’ operates on the five
planes of talk in Shiffrin’s (1987) model (see Section 1.3 for the five planes). Nanka

‘like’ separately or simultaneously operates on the participation framework, expressing

43

the speaker’s hesitancy toward the addressee and/ or uncertainty towards the content of
the utterance. Nanka ‘like’ also operates in the exchange structure and the action
structure proposed by Shiffrin since it occurs turn-initial and it prefaces opposition (1
token 0.4%). Nanka ‘like’ also operates in the information state with its function as a
language-production-based ﬁller. Table 15 summarizes how nanka ‘like’ Operates on

Shiffrin’s ﬁve planes of talk.

Table 15 Summary of nanka ‘like’ on Shiffrin’s five planes of talk

 

 

 

 

 

 

 

ﬁve planes information participation ideational action exchange
state framework structure structure structure
Ifunction secondary rimary secondary secondary I

 

4.4.5. Functions of the filler 4220 ‘well’

The functions of am ‘well’, which was the second most frequently used filler
in the data are examined in this discussion. Ano ‘well’ is originally a demonstrative
adjective meaning ‘that (far away from both the speaker and the hearer)’. Cook
(1993:23) lists the functions of 4220 ‘well’ as follows: to start a turn or a conversation; to
get an the addressee’s attention; to highlight a proposition; to start a new topic; and to
preface a disagreement.

A220 ‘well’ in the present data occurred primarily in the following contexts:
(1) prefacing the main point of utterance; (2) prefacing a new topic ; (3) initiating a turn.
The main function of am ‘well’ is an attention getter, soliciting the addressee’s
cooperation. It has a secondary function as a content softener to avoid abruptness of
the upcoming content. Examples (18), (19) and (20) illustrate three contexts 4220 ‘well’

occurred. First, am ‘well’ is often used prefacing a main point of utterance, 38 tokens

44

(21%), as in example (18) in bold.

(18) S: a. Ee de uchi no syujin jitsu wa anoo kotoshi no
F and my GEN husband actually TOP SF this year GEN

b. natsu de nihon ni kaeru n desu.
summer at Japan to return NOM POL

‘As a matter of fact, well, my husband will return to Japan this summer.’
In (18), the main point of the utterance is that the speaker’s husband is going back to
Japan. A220 ‘well’ in line a functions as an attention getter, signaling the importance of
upcoming content in the utterance in order to solicit the addressee’s cooperation. At
the same time 4220 ‘well’ also functions as a softener in order to reduce embarrassment
when the speaker discusses her personal matter.

Secondly, 4220 ‘well’ is often used prefacing a new topic, 21 tokens (13%), as
shown in example (19) in bold.

(19) S: a. Ano kono mae kara gimon ni omotteta koto ga aru
SF this prior since suspect P think thing SUB have

b. n desu kedo.
NOM COP but

M: c. Hai
yes

‘Well, I have a question I’ve been wanting to ask.’
‘Yes.’

In (19), S starts a new topic prefaced by 4220 ‘well’ in line a. A220 ‘well’ functions to
get an attention and to signal the importance of upcoming content in the utterance in
order to solicit the addressee’s cooperation, as well as to tone down the abruptness of

bringing up the new topic.

45

Finally, 4220 ‘well’ is often used at the beginning of a new speaker’s turn, 17
tokens (9 %). Example (20) illustrates the occurrence of am ‘well’ used as a turn

initiator, shown in bold.

(20) Y: a. Ano ﬁgyasukeeto no senshu ni naritakatta.
SF ﬁgure skate GEN player P become want to

T: b. Sonna koto yutta kke.
such thing say P

‘Well, I wanted to become a figure skater.’
‘Did you tell me that before?’

In (20), Y starts her turn with 4220 ‘well’ in line a. A220 ‘well’ functions as an attention
getter and to signal the importance of upcoming content in the utterance in order to
solicit the addressee’s cooperation, as well as a softener to reduce embarrassment when
the speaker discusses her past experience.

A220 ‘well’ is often accompanied by other attention getting and emphasizing
devices (9 tokens 5%). An attention getting device is any word that has an effect of
soliciting an addressee’s attention, such as hora ‘look’ and an interactional particle 54(4)
‘look’. An emphasizing device is any word that has an effect of emphasizing, such as
sugoi ‘very’ and hontoni ‘really’. Ano ‘well’ never occurred in turn-final position (see
Table 14), which is probably due to the fact that the speaker has no need to get the
addressee’s attention at the end of his/ her turn. These facts support the claim that the
main function of 4220 ‘well’ is to get the addressee’s attention.

In this subsection three contexts were examined in which am ‘well’ was used:
(1) prefacing a main point of utterance; (2) prefacing a new topic ; (3) initiating a turn.

The main function of am ‘well’ is as an attention getter, soliciting the addressee’s

46

cooperation, with a secondary function as a content softener in order to reduce an
abruptness of the upcoming content of the utterance.

The following discussion examines how 4220 ‘well’ operates on the ﬁve planes
of talk in Shiffrin’s (1987) model (see Section 1.3 for the ﬁve planes). Ano ‘well’
operates in the participation framework, getting attention from the addressee and
softening the upcoming content of utterance in order to reduce an abruptness. A220
‘well’ also operates in the exchange structure and the action structure, since it occurs
turn-initial and it prefaces request (2 tokens 1%). A220 ‘well’ also operates in the
information state with its function as a language-production-based ﬁller. Table 16
summarizes how 4220 ‘well’ operates on Shiffrin’s ﬁve planes of talk.

Table 16 Summary of 4220 ‘like’ on Shiffrin’s ﬁve planes of talk

 

 

 

 

 

 

 

ﬁve planes information participation ideational action exchange
state framework structure structure structure
ﬁnaion secondary primary secondary secondary I

 

4.5. Summary

In this chapter, the relation between psychological distance between speakers
and the use of ﬁllers, as well as the functions of the two most commonly used ﬁllers,
nanka ‘like’ and 4220 ‘well’ in Japanese conversation were examined. The ﬁndings are
as follows: Fillers that function both as language-production-based type and socially
motivated type are 4220, ‘well’, nanka ‘like’ et0 ‘well’; some words cannot be
categorized either as fillers or words in other categories; and some ﬁllers cannot be
clearly categorized into language-production-based type or socially motivated type.

Speech styles affect both the frequency and the kind of fillers. More ﬁllers occur in

47

the polite style than in the plain style. The casual sounding filler nanka ‘like’
occurred more frequently in the plain type. Relatively formal sounding ﬁllers, such as
4220 ‘well’ and sono ‘well’, on the other hand, occurred more frequently in the polite
type. This shows that more ﬁller occur when the speaker feels psychologically distant
towards the addressee.

Fillers have important functions as a content softener, showing hesitation
and/ or uncertainty not to be imposing to the addressee, as an attention getter, soliciting
the addressee’s cooperation, and as a softener to reduce an abruptness of the upcoming
content of an utterance. Fillers also function as language-production-based filler,
signaling a repair and intention to continue. It is interpreted that the fillers function
to help to create maximum agreeableness between the speakers.

Fillers 4220 ‘well’ and nanka ‘like’ operate primarily in the participation
framework in Shiffrin’s (1987) model. With functions operating on the other three
planes of talk, the information state, the action structure, and the exchange structure,

ﬁllers contribute to discourse coherence.

48

Chapter 5
Conclusion

The present study investigated a relation between psychological distance
between speakers and the use of ﬁllers, as well as the functions of the two most
commonly used ﬁllers, nanka ‘like’ and 4220 ‘well’, in Japanese conversation. Based on
the data of twelve naturally-occurring conversations, the following conclusions are
made: Fillers are categorized into language-production—based type and socially
motivated type; fillers that function as both types are am ‘well’, nanka ‘like’ and eto
‘well’; some words cannot be categorized either as ﬁllers or words in other categories;
and some ﬁllers cannot be clearly categorized into language-production-based type or
socially motivated type. Speech styles affect both the frequency and the kind of fillers.
More ﬁllers occur in the polite Style than in the plain style. The casual sounding filler
nanka ‘like’ occurred more frequently in the plain type. Relatively formal sounding
ﬁllers, such as 4n0 ‘well’ and sono ‘well’, on the other hand, occurred more frequently
in the polite type. This supports that more ﬁllers occur when the speaker feels
psychologically distant towards the addressee.

The ﬁller nanka ‘like’ has function as a content softener, showing hesitation
and/ or uncertainty not to be imposing to the addressee. The filler 4220 ‘well’ has
functions as an attention getter, soliciting the addressee’s cooperation, and as a softener
to reduce an abruptness of the upcoming content of an utterance. The ﬁllers 4220,
‘well’, nanka ‘like’ and eto ‘well’ function as language-production-based ﬁllers, when a

repair occurs in order to Show the speaker’s intention to continue. It is interpreted

49

from these functions that the fillers help to create maximum agreeableness between the
speakers, as Maynard (1989) optly pointed out.

The ﬁllers nanka ‘like’ and am ’well’ operate primarily in the participation
framework in Shiffrin’s (1987) model. With functions operating on the other three
planes of talk, the information state, the action structure, and the exchange structure,
ﬁllers contribute to discourse coherence.

Future studies may examine the relation between the use and kinds of fillers
and the speaker’s age. In the present study younger participants tended to use more
fillers, but no claims is made in terms of such a relation. Gender comparisons may
also be proﬁtable.

Fillers, especially have not received much attention in the past, since they are
simply regarded as something that just ﬁlls potential silence. However, the present
study shows that they play important roles in structuring discourse coherence in much

the same way as other discourse markers.

50

APPENDIX A

1. Transcription methods

The present study follows the method employed in Maynard’s The japanese
conversation (1989).

2. List of abbreviations

ACC accusative

COP copula

DAT dative

F ﬁller (PF: language-production-based filler SF: socially motivated ﬁller)
GEN genitive

IP interactional particle (often called ‘sentence-ﬁnal particle’)

NEG negative

NOM nominalizer

P particle

PAS Past

PASS passive

POL polite auxiliary verb
Q question particle
QT quotation marker
SUB subject marker
TOP topic marker

51

APPENDIX B

Questionnaire (This will be strictly conﬁdential, so please tell the truth.)

Name: Age:

 

1. Where in Japan are you from?

2. How long have you been living in US?

_1 year or less _2 years or less _3 years or less _4 years or less _5 years or more

3. Are you married? _Yes _No

If Yes, what language do you speak with your spouse?

4. Among all you speak, how much Japanese do you speak a day?

_020% _20-40% _40-60% _60-80% _80100%

5. How long have you known your conversation partner?

_ today is the ﬁrst day _1 year or less (how many months? __mo.)

__2 years or less _3 years or less _4 years or less _5 years or more

6. How would you describe the relationship with your conversation partner?

__ just met _acquaintance _friend _close friend _best friend

7. How did the conversation go? (1- strongly agree <——> 5- strongly disagree)
a. The conversation went smoothly. 1- 2-
b. I was relaxed to talk. 1- 2-
c. I enjoyed talking with my conversation partner. 1- 2-
d. I was nervous in the conversation. I- 2-
e. I was uncomfortable in the talk. 1- 2-

1- 2-

f. The atmosphere of the conversation was not pleasant.

8. Is there anything you would like to add to the above with regard to how you felt
about the conversation?

52

APPENDIX C

CONSENT FORM

Explanation of the project

This is linguistic research on conversation in Japanese. You will be asked to
participate in a 10-15 minute conversation with your conversation partner assigned by
the researcher. This conversation will take place anywhere convenient for you.
Japanese language will be used. The conversation will be recorded on a cassette tape
and later may be used as data for my master’s thesis. The topic of conversation will be
assigned by the researcher. However you are fully entitled to choose what to talk and
what not to talk on the topic. No part of the conversation will be personally
intrusive in nature. All results will be treated with strict conﬁdence and you will
remain anonymous in any report of research ﬁndings. On request and within these
restrictions results will be made available to you. In accordance with Michigan State
University policy on research involving human subjects, the conversation will not be
harmful to you in any way. The participation is voluntary and you may withdraw at

any time without penalty.

If you agree to participate in this project, please Sign and date on the lines below.
If you have any concerns regarding this project, please contact Yoshiko Yamamoto at

1537 B Spartan Village, East Lansing, MI 48823, (517) 355-2957.

Thank you very much for your cooperation.

 

I have read the explanation of the project above and understand its content. I
agree to participate in this project.
Name (print)

 

Signature Date

 

53

BIBLIOGRAPHY

Bloch, Charlotte, 1996. ‘Emotions and discourse’ Text Vol. 16, No. 3, 323-341

Brown, Gillian and Yule, George, 1983. ‘Discourse analysis’ Cambridge University
Press, Cambridge, England.

Fraser, Bruse, To appear. ‘What are discourse markers?’ Boston University.

Cook, Haruko, M, 1993. ‘Functions of the ﬁller 4220 in Japanese’ japanese Korean
linguistics Vol. 2 CLSI Publications, California.

Holms, Janet, 1986. ‘Functions of you know in women’s and men’s speech’ Language in
society Vol. 15, 1-22

Hudson, Mutsuko Endo, 1998. ‘So? (on Japanese connectives sorede, dakara, and ja.)’
[apanese Korean Linguistics Vol. 7 CLSI Publications, California.

Ito, Rika, 1995. ‘The functions of the Japanese conjunction (sore)de as a discourse
marker’ M.A. thesis. Michigan State University.

Kamio, Akio, 1990. ‘Joohoo no nawabari riron, Territory of Information’ Taisyukan
Syoten, Tokyo.

Kinsui, Satoshi, 1990. ‘Shijishi to danwa no kozo’ Gengo Vol. 19 No. 4 Taisyukan
Syoten, Tokyo.

Kuno, Susumu, 1973. ‘The Structure of the Japanese language’ The MIT Press,
Cambridge, Massachusetts.

Levinson, Stephen, C., 1983. ‘Pragmatics’ Cambridge University Press, Cambridge,
England.

Makino, Seiichi and Tsutsui, Michio, 1986. ‘A dictionary of basic Japanese grammar’

54

The Japan Times, Tokyo, Japan

Maynard, Senko, Kumiya, 1989. ‘Japanese conversation’ Ablex Publishing, Norwood,
New Jersey.

Onodera, Noriko Okada, 1993. ‘Pragmatic change in Japanese: Conjunctions and

interjections as discourse markers’ Unpublished dissertation. Georgetown
University.

Philips, Mieko, 1998. ‘An aspect of conversation management in Japanese: Use of
connectives, ﬁllers and interactional particles as discourse markers’

Unpublished dissertation. Michigan State University.

Sacks, Harvey, Schegloff, Emanuel, A., and Jefferson, Gail, 1978. ‘A simplest
systematics for the organization of turn taking for conversation’ in Jim
Schenkein (Ed..) Studies in the organization of conversational Interaction
Academic Press, New York.

Owen, Marion, 1981. ‘Conversational units and the use of ‘well...’ in Paul Werth (Ed..)

Conversation and discourse Croom Helm, London.

Saito, Makiko, 1992. ‘A Study of the discourse marker nanka in Japanese’ M.A. thesis.
Michigan State University.

Schiffrin, Deborah, 1987. ‘Discourse markers’ Cambridge University Press, Cambridge,
England.

Shourup, Lawrence, C, 1982. ‘Common Discourse Particles in English Conversation’
Ph.D. dissertation, Ohio State University.

Shibatani, Masayoshi, 1990. ‘The languages of Japan’ Cambridge University Press,
Cambridge, England.

Stubbs, Michael, 1983. ‘Discourse analysis’ Basil Blackwell Publisher, Oxford,
England.

55

White, Sheida, 1989. ‘Backchannels across cultures: A study of Americans and
Japanese’ Language in Society Vol. 18, 59-76.

Yamamoto, Yoshiko, 1997. ‘Functions of the ﬁller 4220 and 50220 in Japanese’,

unpublished manuscript.

56

"I11111111113111111ES