 

5.. .. a .
I 1.2:“:
.r . :33.
3.3.3.3};
.

.mh.

. 3 1:.)57 h!

m.

m: .fr a.

h
(#5. 3i? 2
Eyrm‘wu .1 .3? ..

. .3.
: - a
.n. 1.. 2.1st

I". .3.

4 Jo .. a «ﬁg.
veﬁrﬁabﬁﬁuﬁw

.. , .5 5-1:.
ngAﬁw .3933. . X.

#9.;
(3.".
xicsruy. 3..er .
.5 li£ﬂ~lli~ $1 snot}...
5 “but FEHMVﬁFI. {’F.. 04”.},
1. Jun! ! .5 but;
. . it: a. lam-Mum":
I... .
9%Jﬁi
ﬂqupruw .9 Mn
iv;ﬁ.ta _ p

: v.91..:.\f
)i. It:
.-

 

3.5 n

.t ”v.73... .. . .

3
Ir.

TERMS

(3:) f r 3
51/1. #11320

 

 

LIBRARY

University

Michigan State

 

 

This is to certify that the

dissertation entitled

THIS GIRL WANTS OUTI: AN ANALYSIS OF
NEED/NANT+PREP6$ITIONAL PHRASE AND NEED/NANT+
PREPOSITIONAL ADVERB

presented by

ERICA JUNE BENSON

has been accepted towards fulﬁllment

of the requirements for

Ph.D. degree in LINGUISTICS

 

 

 

Date i// 91/0 3

MS U is an Afﬁrmative Action/Equal Opportunity Institution

Major professor

042771

..___ ,___
-~---~ ' V

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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRCIDateDue.p65cp. 15

THIS GIRL WANTS OUTI: AN ANALYSIS OF NEED/WANT + PREPOSITIONAL
PHRASE AND NEED/WANT + PREPOSITIONAL ADVERB

By

Erica June Benson

A DISSERTATION

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

DOCTOR OF PHILOSOPY
The Department of Linguistics and Germanic, Slavic, Asian and African Languages

2003

ABSTRACT
THIS GIRL WANTS OUTI: AN ANALYSIS OF NEED/WANT + PREPOSITIONAL
PHRASE AND NEED/WANT + PREPOSITIONAL ADVERB
By

Erica June Benson

The aims of this dissertation are to examine the geographic, social, and stylistic
distributions of the grammatical constructions need/want + prepositional phrase (e. g.
Dean needs in the building) and need/want + prepositional adverb (e. g. The cat wants
out) and to investigate other factors bearing on these distributions, e. g. the semantic
features of concrete and abstract. Most previous literature on these constructions has
looked at want + prepositional adverb; almost no literature has mentioned need/want +
prepositional phrase and need + prepositional adverb.

Several ﬁndings have emerged from questionnaire data gathered ﬁ'om 163
respondents, in Ohio, Michigan, and Georgia. All of these constructions show robust
Midland distributions, and although need/want + prepositional phrase and need +
prepositional adverb were rarely (or never) mentioned in other sources, use of them was
reported by over 75% of respondents in the Midland. Furthermore, these constructions
appear to be used more than previous literature had suggested in the North and the South,
where some forms, e. g. want down, want in, want oﬁi were reported by more than 60% of
respondents.

More importantly, the social and stylistic distributions of these constructions in

the Midland challenges long held assumptions about the relationship between these two

variables. Need/want + prepositional phrase and need/want + prepositional adverb show
little to no variation across the social variables of urban-rural residence, gender,
socioeconomic status, and age. These constructions are, however, stratiﬁed by style.
Although many respondents report using need/want + prepositional phrase and need/want
+ prepositional adverb in all styles, in general, these constructions are much more
frequent in informal styles than in formal styles. Stylistic variation without social
variation calls into question not only the assumption that social and stylistic variation are
inherently related (e.g. Romaine 1980; Labov 1972; Bell 1984; Finegan & Biber 1994)
but also the idea that the amount of stylistic variation cannot exceed that of social
variation, i.e. Bell’s style axiom (1984).

This study of need/want + prepositional phrase and need/want + prepositional
adverb reveals their robust distribution in the Midland dialect area, which together with
historical evidence, supports the idea of a distinct Midland dialect. In addition, use in the
North and South is greater than expected, which would appear to be the result of spread
of these constructions from the Midland. The ﬁndings also challenge long-held
assumptions about the interrelated nature of social and stylistic distribution. Furthermore,
this study of grammatical features raises questions about general characteristics of
language variation and change, which have largely been based on studies of

phonetic/phonological variables.

Copyright by
ERICA JUNE BENSON
2003

Dedicated to the wise ones in my family who taught me
more about life and learning than they know

ACKNOWLEDGMENTS

First and foremost, I wish to thank my chair, Dennis R. Preston, for being a great role
model and for giving unselﬁshly of his time and wisdom, not only during the dissertation
writing process but also throughout my graduate studies at Michigan State. I would also
like to express my gratitude to my other committee members, Barbara Abbott, Grover
Hudson, and David Dwyer for their beneﬁcent guidance and insightful feedback.

I would like to acknowledge those who completed surveys for this project and
especially those respondents who generously devoted their time to be interviewed. I also
wish to recognize several colleagues who distributed questionnaires to their students,
including Bridget Anderson, Gord Easson, Peggy Ketner, Nevin Leder, and Patrick
Mullen.

I am indebted to countless friends, fellow students, and family members who saw
me through this endeavor. A special thanks to the members of the sociolinguistics gang,
past and present, who discussed my project (on numerous occasions) and provided
invaluable feedback, including Nevin, Betsy, Terumi, Chunhua, Becky, Githinji, Mike,
Bartek, Cathy, Paul, Yuliya, Lille-Kristen, Jill, and others. I am especially grateful to
Robin Pulford, whose critical comments and constructive advice contributed much to this
dissertation. Finally, I am very thankful for the support of my family and my close
friends, including Valorie Simmons, Tomomi Kakegawa, Terumi Imai, Betsy Evans,
Chunhua Ma, Jane Ozanich, Gabriela Alfaraz, Nancy Millenbaugh, Maria McCaffrey,

and Alexxis Avalon.

vi

TABLE OF CONTENTS

LIST OF TABLES ........................................ ix
LIST OF FIGURES .................................................................................................. xii
KEY TO ABBREVIATIONS .................................................................................. xiii
CHAPTER 1

1 . Introduction .................................................................................................... 1
CHAPTER 2

2. Background .................................................................................................... 5
2.1. Syntactic properties of the verbs need and want .............................................. 5
2.2. Literature overview ........................................................................................ 11
2.2.1. N/W+PP ........................................................................................................ 14
2.2.2. N/W+PA ....................................................................................................... 16
2.3. Geographic distribution ................................................................................. 21
2.3.1. The Midland .................................................................................................. 23
2.3.2. The South ...................................................................................................... 29
2.3.3. The North ...................................................................................................... 30
2.3.4. Geographic expansion .................................................................................... 32
2.4. Semantic features ........................................................................................... 35
2.5. Social factors ......................................................................... ' ........................ 38
2.5.1. Rural-urban residence .................................................................................... 38
2.5.2. Social class .................................................................................................... 41
2.5.3. Age ................................................................................................................ 43
2.5.4. Gender ........................................................................................................... 44
2.6. Style .............................................................................................................. 46
2.7. Hypotheses .................................................................................................... 51
CHAPTER 3

3. Methodology ................................................................................................. 53
3.1. Linguistic variables ........................................................................................ 53
3.2. Survey instruments ........................................................................................ 56
3.2.]. Study of grammatical features ........................................................................ 56
3.2.2. Questionnaires ............................................................................................... 57
3.3. Respondents and nonlinguistic variables ........................................................ 59
3.4. Data manipulation and statistical analyses ..................................................... 64
CHAPTER 4

4. Results and discussion ................................................................................... 66
4.1. Use across dialect regions .............................................................................. 66
4.1.1. N/W+PP ........................................................................................................ 67

vii

4.1.2. N/W+PA ....................................................................................................... 70
4.1.3. Discussion ..................................................................................................... 75
4.2. Semantic features ........................................................................................... 81
4.2.1. N/W+PP ........................................................................................................ 81
4.2.2. N+PA ................................................................... _. ........................................ 84
4.2.3. W+PA ........................................................................................................... 87
4.2.4. Discussion ..................................................................................................... 90
4.3. Usage hierarchies ........................................................................................... 95
4.3.]. Need vs. want ............................................................................................... 95
4.3.2. N/W+PP vs. N/W+PA ................................................................................. 101
4.3.5. Discussion ................................................................................................... 103
4.4. Nonlinguistic factors .................................................................................... 105
11.41.] . ‘Ititrtil-tiriaziri ................................................................................................. il()f$
4.4.2. Gender ......................................................................................................... 108
4.4.3. Socioeconomic status ................................................................................... 112
4.4.4. Age .............................................................................................................. 114
4.4.5. Discussion ................................................................................................... 1 16
4.5. Stylistic variation ......................................................................................... 125
4.5.1. N/W+PP ...................................................................................................... 126
4.5.2. N+PA .......................................................................................................... 129
4.5.3. W+PA ......................................................................................................... 131
4.5.3. Discussion ................................................................................................... 133
CHAPTERS

5. Conclusion .................................................................................................. 139
5.1. ‘ Summary of ﬁndings ............................................................. ' ...................... 139
5.2. Limitations and suggestions for future research ........................................... 141
5.3. Relevance of ﬁndings to ﬁeld of linguistics ................................................. 143
APPENDICES

Appendix A: Questionnaire items ............................................................................ 149
Appendix B: Sample questionnaire .......................................................................... 155
Appendix C: Essential information on dialectology projects .................................... 160
Appendix D: List of respondents ............................................................................. 163
Appendix E: Learning curve statistics ..................................................................... 169
REFERENCES ....................................................................................................... 170

viii

LIST OF TABLES

Table 2.1. Dialect word list reports of W+PA in the Midland .................................... 28
Table 2.2. Dialect word list reports of W+PA in the South ........................................ 29
Table 2.3. Use of need/want/like + past participle by socioeconomic class ................ 42
Table 2.4. Use of need/want/Iike + past participle by age ........................................... 44
Table 2.5. Use of want off and need/want/like + past participle by gender .................. 46
Table 3.1. Demographic distribution of respondents .................................................. 60
Table 3.2. Occupational prestige ratings .................................................................... 63
Table 4.1. Use of N/W+PP by dialect region ............................................................. 69
Table 4.2. Use of N+PA by dialect region ................................................................. 72
Table 4.3. Use of W+PA by dialect region ................................................................ 73
Table 4.4. Rank order of N/W+PP and N/W+PA ............................... . ........................ 79
Table 4.5. Concrete and abstract senses of N/W+PP across dialect regions ................ 82
Table 4.6. Concrete and abstract senses of N/W+PP within dialect regions ................ 83
Table 4.7. Concrete and abstract senses of N+PA across dialect regions .................... 85
Table 4.8. Concrete and abstract senses of N+PA within dialect regions .................... 86
Table 4.9. Concrete and abstract senses of W+PA across dialect regions ................... 88
Table 4.10. Concrete and abstract senses of W+PA within dialect regions ................. 89

Table 4.11. Rank order of concrete and abstract senses N/W+PP and N/W+PA ......... 92

Table 4.12. Use of N+PP vs. W+PP .......................................................................... 96
Table 4.13. Use of concrete and abstract senses of N+PP vs. W+PP .......................... 97
Table 4.14. Use of N+PA vs. W+PA ......................................................................... 99

ix

Table 4.15. Use of concrete and abstract senses of N+PA vs. W+PA ....................... 101

Table 4.16. General use of PP vs. PA ...................................................................... 102
Table 4.17. Use of PP vs. PA ............................................ , ....................................... 103
Table 4.18. Use of N/W+PP by rural-urban residence .............................................. 106
Table 4.19. Use of N+PA by rural-urban residence .................................................. 107
Table 4.20. Use of W+PA by rural-urban residence ................................................. 108
Table 4.21. Use of N/W+PP by men and women ..................................................... 109
Table 4.22. Use of N+PA by men and women ......................................................... 1 10
Table 4.23. Use of W+PA by men and women ........................................................ 111
Table 4.24. Use of N/W+PP by socioeconomic status .............................................. 113
Table 4.25. Use of N/W+PA by socioeconomic status ............................................. 113
Table 4.26. Use of N/W+PP by age ......................................................................... 115
Table 4.27. Use of N/W +PA by age ........................................................................ 116
Table 4.28. Use of all tokens of N/W+PP by style ................................................... 127
Table 4.29. Use of all tokens of N+PA by style ....................................................... 130
Table 4.30. Use of all tokens of W+PA by style ...................................................... 132
Table 5.1. Effect of animacy/agency of the subject of need/want on distribution ..... 144
Table A. l. Token sentences of N/W+PP .................................................................. 149
Table A2. Token sentences of N+PA ...................................................................... 150
Table A3. Token sentences of W+PA ..................................................................... 151
Table A.4. Filler sentences ........................................................................... 152
Table A5. Widely accepted ﬁller sentences ............................................................ 153
Table C. 1 . Linguistic Atlas of New England (LANE) ................................................ 160

Table C.2. Linguistic Atlas of the Middle and South Atlantic States (LAMSAS) ........ 160

Table C.3. Linguistic Atlas of the North Central States (LANCS) ............................. 160
Table c.4. Linguistic Atlas ofthe Upper Midwest (LA UM .. .................................... 160
Table C.5. Linguistic Atlas of the Gulf States (LA GS) .............................................. 160
Table C.6. Linguistic Atlas of the Pacific Northwest (LAPNW) ................................ 161
Table C.7. Dictionary of American Regional English (DARE) ................................. 161
Table D. 1. List of respondents ................................................................................. 163

Table E. 1. Use of token sentences in three groupings corresponding to the beginning,
middle, and end of the questionnaire ....................................................................... 169

xi

LIST OF FIGURES

Figure 2.1. United States dialect areas .............................. _ ......................................... 22
Figure 2.2. W+PA in Linguistic Atlas project data ..................................................... 25
Figure 2.3. W+PA in DARE data ............................................................................... 26
Figure 2.4. Use of want oﬂ in DARE by community type .......................................... 39

Figure 2.5. Socioeconomic and style stratiﬁcation of (ing) for European-American
adults in New York City ............................................................................................ 48

Figure 2.6. Hypothetical case of socioeconomic stratiﬁcation with no stylistic
variation .................................................................................................................... 49

Figure 2.7. Socioeconomic and stylistic stratiﬁcation with hypercorrection of (r) in

adults in New York City ............................................................................................ 50
Figure 4.1. Use of N+PP and W+PP .......................................................................... 68
Figure 4.2. Use of N+PA and W+PA ......................................................................... 71
Figure 4.3. Use ofN+PP vs. W+PP ........................ 95
Figure 4.4. Use of N+PA vs. W+PA .......................................................................... 98
Figure 4.5. Use of N/W+PP and NfW+PA by style .................................................. 134
Figure 4.6. W+PP by socioeconomic class and style ................................................ 137

xii

Ade
AH
DARE
DP
LAGS
LAMSAS
LANCS
LANE
LAPNW
LA UM
N+PA
NP
N/W+PA
N+PP
N/W+PP
OADLG
OED
PA

PP

RH
W+PA
W+PP
VP

KEY TO ABBREVIATIONS

Adjective phrase

American Heritage College Dictionary
Dictionary of American Regional English
determiner phrase

Linguistic Atlas of the Gulf States
Linguistic Atlas of the Middle and South Atlantic States
Linguistic Atlas of the North Central States
Linguistic Atlas of New England
Linguistic Atlas of the Paciﬁc Northwest
Linguistic Atlas of the Upper Midwest
need + prepositional adverb

noun phrase

need/want + prepositional adverb

need + prepositional phrase

need/want + prepositional phrase

Oxford American Dictionary and Language Guide
Oxford English Dictionary

prepositional adverb

prepositional phrase

Random House Dictionary

want + prepositional adverb

want + prepositional phrase

verb phrase

xiii

1. Introduction

In all varieties of English, the following sentences1 are grammatical:

( 1) I’m sorry to bother you, but I need to get by.

(2) The baby wants to get down.

(3) When the dog scratches at the back door, he wants to come in.
(4) The ﬁlmes from this bus are making me sick. I need to get off.
(5) This box needs to go/to be put in the car.

(6) I want to get off this horse.

In these sentences and others like them, the verbs need and want are followed by the
inﬁnitive form of a verb and a prepositional adverb, as in 1-4, or a prepositional phrase,

as in 5-6. In some varieties of English, however, another form is acceptable:

(7) I’m sorry to bother you, but I need by. (17)2

(8) The baby wants down. (3 6)

(9) When the dog scratches at the back door, he wants in. (3 8)
(10) The fumes from this bus are making me sick. I need off. (25)
(l 1) This box needs in the car. (42)

(12) I want off this horse. (l3)

 

] Three types of examples are used in this study:

(i) examples invented to illustrate grammatical points;

(ii) examples invented for use on the questionnaire, listed in Appendix A (see Footnote 2);

(iii) examples borrowed from other sources and appropriately cited.

2 Following type ii examples (see Footnote 1) is the number of the questionnaire item in Appendix A.

The constructions of concern to this study are need/want + prepositional adverb,
henceforth N/W+PA, as in 7-10, and need/Want + prepositional phrase, henceforth
N/W-rPP, as in 11-12. The choice of the term ‘prepositional phrase’ for the constructions
in 11-12 is straightforward. The term ‘prepositional adverb’ for the words following
need/Want in 7-10 requires some explanation. In the literature, several terms have been
used to refer to what I am calling ‘prepositional adverbs’, including ‘adverb’ (Griebling
1947 :5), ‘adverb of place’ (Ashcom 1953:255; Cassidy & Hall 1996:769), ‘directional
adverb’ (Webster ’s, 3rd ed., s.v. 2 ‘want’ 5; Crozier 1984), ‘obligatory adverbial’ (Quirk
et al 19852732), ‘particle’ (Bumham 19262395; Marckwardt 1948), and ‘preposition’
(Maxfreld 193120; A. Wilson 1948:238; Griebling 1947:5). Although the term
‘prepositional adverb’ does not appear in other previous sources, it seems like an
appropriate label for these words for two reasons. One, these words look like prepositions
(even though they do not behave like prepositions, since they are not followed by noun
phrases), and in other cases, they can take noun phrase complements. Two, by using the
term ‘prepositional adverb’ a relationship between bare forms (i.e. prepositional adverb
forms) and prepositional phrase constructions is evident. The addition of the term
‘adverb’ not only acknowledges that these words are not structurally prepositions (since
no noun phrase follows) but also indicates that these words are like some others that do
not fit nicely into currently accepted word categories. I, therefore, adopt the term
‘prepositional adverb’ but recognize that the exact syntactic category of these words
remains to be determined.

Currently, N/W+PP and N/W+PA are not well understood with respect to their

geographic, social, and stylistic distribution in the United States. Remarkably little

attention has been paid to these constructions by linguists and dialectologists. In spite of
the fact that there are numerous pronouncements about some forms of these constructions
as regional, informal, or even nonstandard, there are few studies that have systematically
investigated even one form.

My desire to work with these constructions arose from my interests in three areas:
the debate over the status of the Midland dialect area; the role of grammatical variables in
the ﬁeld of language variation and change; and the need for more detailed study of
linguistic and nonlinguistic factors that affect the geographic, social, and stylistic
distribution of grammatical variables.

The research aims of this study include the following:

0 To examine the distribution of N/W+PP and N/W+PA across different regional and
social groups; that is, to gain some insight into how these forms pattern across
different dialect areas, rural-urban areas, socioeconomic classes, ages, and genders.

0 To explore the role of the semantic features of concrete and abstract in the
distribution of these constructions.

0 To investigate differences in the use of various forms of N/W+PP and N/W+PA,
including differences between need and want, PP and PA, and various prepositions.

0 To examine the use of N/W+PP and N/W+PA across different situations of use, i.e.

style (or register).

In Chapter 2, I review the literature on N/W+PP and N/W+PA, primarily from

dialect word lists and dialectology projects as well as dictionaries and usage guides, and I

close with a list of hypotheses stemming from the previous literature. Chapter 3 describes
the methodology for this project: the linguistic variables, the social variables, the
respondents, the data collection methods, and the statistical analyses used on the data.
Chapter 4 presents the results and discussion. Finally, Chapter 5 includes concluding

remarks and suggestions for further research.

2. Background

The goal of this section is to review the previous literature on the constructions N/W+PP
and N/W+PA. This review will demonstrate that they are more than deserving of
attention in linguistic research in the United States and will allow me to propose
hypotheses with respect to their geographic, social, and stylistic distribution. Section 2.1
presents an overview of the syntactic properties of the verbs need and want with an
emphasis on N/W+PP and N/W+PA constructions, though parallels are also drawn
between them and other construction with need/want. In Section 2.2, the focus is on the
various prepositional phrases and prepositional adverbs found to occur in N/W+PP and
N/W+PA constructions. Section 2.3 examines the geographic distribution of N/W+PP
and N/W+PA in the United States, which has been variously described as ‘regional’,
‘Midland’, ‘Southern’, and ‘American’. Section 2.4 discusses the semantic features of
concrete and abstract and the potential relevance of these features to the use and
acceptability of N/W+PP and N/W+PA. In Section 2.5, social factors that have been
shown to play a role in studies of language variation are examined, e.g. rural-urban
residence, socioeconomic status, age, and gender. Section 2.6, reviews what is known
about the use of N/W+PP and N/W+PA in different styles and explores the relationship
between stylistic variation and social variation. Finally, Section 2.7 lists the hypotheses

for the present study.

2.1. Syntactic properties of the verbs need and want
The purpose of this section is to outline some of the basic syntactic properties of the

verbs need and want, particularly as they relate to the analysis of N/W+PP and N/W+PA

and to identify some potentially related constructions. This section is not intended as a
complete syntactic analysis; instead, it is intended to provide the descriptive basis for the
purpose of a dialectological and sociolinguistic analysis. In all varieties of American
English, need and want can take determiner phrase complements (DP), as in 1-2 but not

adjective phrase complements (Ade), as in 3, or verb phrase complements (VP), as in 4.

(1) The child needs/wants candy.
(2) The child needs/wants more candy.
(3) *The child needs/wants happy.

(4) *The child needs/wants tell a story.

Another common complement of need/want is the nonﬁnite clause. Of particular
interest here are nonﬁnite clause complements consisting of an empty (PRO) subject and
a to- inﬁnitive. In such cases, the PRO subject of the to-inﬁnitive gets its interpretation

from the subject of need/want, as in 5-10.

(5) The child needs/wants to have candy.

(6) The child needs/wants to have more candy.
(7) The child needs/wants to be happy.

(8) The child needs/wants to tell a story.

(9) The child needs/wants to come in.

(10) The child needs/wants to come in the house.

It is possible to passivize the to-inﬁnitive verb, as in 11-12. In these cases,
however, the PRO subject of the to-inﬁnitive, which still gets its interpretation from the
subject of need/want, originates as an argument of the inﬁnitive verb, from which it gets

its theta-role assignment.

(11) The child needs/wants to be told a story.

(12) The house needs to be painted.

A similar passivization effect is achieved when need/want is followed by a nonﬁnite
clause with an empty (PRO) subject and a gerund; that is, the interpretive relationship
outlined previously holds between the subject of need/want and the PRO subject of the

gerund, which originates as an argument of the gerund, as shown in 13-14’2.

(1 3) The shoes need/want polishing.

(14) The house needs painting.

 

lSelectional restrictions may limit, to some extent, the possibilities, though they are not crucial to the
argument here. In the constructions in which need and want are followed by a to-inﬁnitive, the inﬁnitive
verb has a PRO subject that gets its interpretation from the subject of need and want, and this inﬁnitive
verb imposes selectional restrictions on the subject of need and want. So, sentence i is ungrammatical
because the selectional restriction of an animate subject for paint is not met, but ii is grammatical because
bake can take an inanimate subject.

(i) *The house needs/wants to paint.

(ii) The cake needs to bake for 45 minutes.

2In addition, there is a selectional restriction on the subject of want + nonﬁnite clause; namely, its subject
must be animate. In some dialects of English, the subject of want may be inanimate; however, in those
cases, want is understood as need. So, 14, 16, and 18 are grammatical in all varieties of English, while iii-v
are grammatical in only some varieties.

(iii) The house wants to be painted.

(iv) The house wants painting.

(v) The cake wants to bake for 45 minutes.

Need/want may be not only followed by nonﬁnite clauses with empty subjects but

also nonﬁnite clauses with overt subjects, as in 15-20.

(15) The child needs/wants you to have candy.

(16) The child needs/wants her to have more candy.
(17) The child needs/wants his friend to be happy.

(1 8) The child needs/wants John to tell a story.

(19) The child needs/wants her babysitter to come in.

(20) The child needs/wants the dog to come in the house.

The structural analysis of these constructions is certainly important for syntactic theory
and perhaps for other sociolinguistic endeavors; however, it is beyond the scope of this
study for two reasons. First, the syntactic properties of clauses with overt subjects are
different from those with PRO subjects; for example, in sentences with nonﬁnite clauses
with empty (PRO) subjects, another element in the sentence is linked to the PRO;
however, in nonﬁnite clauses with overt subjects, no such dependency relationship exists.
Second, adding clauses with overt subjects to the pool of token sentences would have
made the survey instruments prohibitively large and cmnbersome for respondents;
therefore, the analysis of need and want followed by clauses with overt subjects and the
comparison to clauses with empty subjects are left to be studied in a future project.
Returning to the empty subject constructions that are central to this study, in
addition to nonﬁnite clause complements, in some varieties of American (and Canadian

and Scotch-Irish) English, need and want may take other complements. Of particular

interest to this study is that the inﬁnitive verb (to + V) is not always obligatory with a
PRO subject, and in some varieties of English sentences, like 21-24 (compare with 11-12

and 9-10, respectively) are well-formed.

(21) The child needs/wants told a story.
(22) The house needs painted.
(23) The child needs/wants in.

(24) The child needs/wants in the house.

On ﬁrst reading, sentences like 21-24, may seem odd to speakers whose dialects do not
allow these constructions; such speakers may consider these constructions at best
nonstandard and at worst nonnative (K. Wilson 1993; see quotations in Frazer et a1 1996).
Sentences 21-22 and other constructions involving need/want + past participle are
indeed of great interest to the ﬁeld of language variation and have recently received much
deserved attention in a series of articles in American Speech (Frazer et al. 1996; Murray
& Simon 1999, 2002). In this study, the focus is on constructions like 23-24 in which
need and want are followed by prepositional phrases or prepositional adverbs. On the

surface, these sentences appear to be similar to 1-2 (repeated here).

(1) The child needs/wants candy.

(2) The child needs/wants more candy.

Like 1-2, sentences 23-24 have a DP subject (the child), verb (need or want), and a
complement immediately following the verb (candy, more candy, in, and in the house).
Like in sentences 1-2, in sentences 23-24, the complement of the verb may be preposed.
Furthermore, just as 1-2 are considered to be alternative ways of expressing sentences 5-6
(repeated here), respectively, sentences 23-24 may be considered alternative ways of

expressing 9-10 (repeated here), respectively.

(5) The child needs/wants to have candy.
(6) The child needs/wants to have more candy.
(9) The child needs/wants to come in.

(10) The child needs/wants to come in the house.

This cursory look at the syntactic properties of N/W+PP and N/W+PA raises
more questions than it answers. Indeed, much remains to be done in order to complete a
thorough syntactic analysis of N/W+PP and N/W+PA. What is the syntactic structure of
these constructions? Are N/W+PP and N/W+PA related to need/want + past participle?
Do the constructions N/W+PP and N/W+PA share more than superﬁcial similarities to
N/W+DP? Is there a relationship between the constructions with to-inﬁnitives and those
without to-inﬁnitives? The answers to these questions and others are important to a
comprehensive syntactic analysis of the English language in all its variation but are left to
qualiﬁed syntacticians. For the purposes of this study, the descriptive outline given above
will sufﬁce. F urtherrnore, N/W+PP and N/W+PA, seem to have parallels to other

regionally distributed constructions, e. g. need/want + past participle, as well as to widely

10

accepted constructions, e. g. N/W+DP. The following sections review the previous

literature on N/W+PP and N/W+PA.

2.2. Literature overview
This section outlines the types of literature referencing N/W+PP and N/W+PA,
introduces the range of prepositional phrases and prepositional adverbs that may occur
with need and want, and examines the treatment of N/W+PP and N/W+PA in the
literature. The majority of published work with some mention of N/W+PP and N/W+PA
can be classiﬁed into one of three categories: usage guides, dialect word lists, and
dialectology projects and studies.

Although not typically used in linguistic studies, usage guides, i.e. dictionaries
and grammar books, provide an array of information that may be useful in the study of a
language phenomenon. The mere presence or absence of a linguistic‘item in usage guides
says something. Moreover, usage guides not only provide examples of use of linguistic
items but also often give information about their geographic and stylistic distribution.
Furthermore, as is the case here, usage guides may even identify semantic distinctions
that are not found in other sources, as discussed in Section 2.6. The major weaknesses of
usage guides include the following: 1) the lack of quoted references on which
generalizations about the usage and the distribution of linguistic items are based and 2)
their prescriptive nature, which may result in a commonly used item not being included at
all or given only cursory treatment.

Dialect word lists and reports are included in the second category of published

works that mention N/W+PP and N/W+PA. Most of the dialect word lists cited in this

11

study were based on a single individual’s observations in a particular locale and were
published in the ﬁrst half of the twentieth century in journals like Dialect Notes and
American Speech. The major value of dialect word lists is that they give some indication
of the geographic distribution of the constructions in the early twentieth century, if one
assumes that the observations are indeed representative of the region. Some dialect word
lists are noteworthy for the range of prepositional adverbs and prepositional phrases cited
for a particular area. Unfortunately, dialect word lists rarely include information about the
numbers and characteristics of speakers or about the situations in which such forms are
used; therefore, there is little to be found in dialect word lists about the social distribution
and stylistic distribution.

Dialectology projects and studies make up the third group of published works that
offer data on N/W+PP and N/W+PA. The dialectology projects cited in this study include
the Linguistic Atlas of New England (LANE), the Linguistic Atlas of the Middle and South
Atlantic States (LAMSAS), the Linguistic Atlas of the North Central States (LANCS), the
Linguistic Atlas of the Upper Midwest (LA U110, and the Linguistic Atlas of the Gulf
States (LAGS), collectively referred to as the Linguistic Atlas projects, as well as the
Dictionary of American Regional English (DARE) (see Appendix C). Also included in
this group are studies that either used the data (e.g. Shuy 1962; Dakin 1966a, b; Allen
1975) or adopted the methodology and questionnaire items (e.g. Hartman 1966; Foley
1972) of the Linguistic Atlas projects. Through extensive oral interviews, the Linguistic
Atlas projects and DARE collected data on a variety of linguistic items, the majority
involving phonetic or lexical differences. Trained ﬁeldworkers used worksheets with

prompts that were designed to elicit the linguistic item of interest. Responses were

12

usually recorded manually in phonetic transcription; however, in some cases, e. g. IAGS,
responses were tape recorded and later transcribed.

Data from dialectology projects offer several advantages over other sources. The
data from the Linguistic Atlas projects and DARE were systematically gathered using
trained ﬁeldworkers over large geographical regions. Within these regions, the number of
respondents and some social information, e. g. age and level of education, were known.
There are, however, some drawbacks to using these data.

The data from these dialectology projects are now 20 to 72 years old; the data for
LANE were collected in 1931-1933; LAMSAS, 1933-1974; LANCS, 1933-1978; LAUM,
1949-1962; LAGS, 1968-1983; and DARE, 1965-1970 (see Appendix C). Although these
data may offer some historical perspectives, they cannot, in any way, be construed to
represent current use or use at one particular time in the past. Moreover, while the data
attempt to provide an overview of language use in a particular region, they come from a
small number of respondents that are presumed to represent thousands of actual speakers,
and especially in the early Linguistic Atlas projects, the respondents were predominately
older, rural, males (the ‘NORMS’, nonmobile, older, rural males, of Chambers &
Trudgill 1980:33). Furthermore, the interviewing skill and recording accuracy of
ﬁeldworkers varied widely (see e.g. Thomas 1993:262; Bailey & Tillery 1999). Perhaps
most important to this study is that no forms of N/W+PP and N+PA and very few forms
of W+PA were actually examined. Only want of was included in the questionnaires of
the Linguistic Atlas projects and DARE. In addition, want in and want out were included
at the discretion of ﬁeldworkers in LANCS and LA UM (Allen 1975:72), as was want out

in LAGS (Pederson 1991); other occasional uses outside the questionnaire interview itself

13

were recorded as well. In spite of these weaknesses, dialectology projects are an
important source of data on W+PA, particularly since they provide the only
systematically collected data.

In the remainder of this chapter, material and data from usage guides, dialect
words lists, and dialectology projects inform current understanding of the constructions
N/W+PP and NIW+PA. The bulk of the background literature references only one of
these constructions: W+PA; consequently, the majority of this chapter is concerned with
W+PA. In light of the abundant literature on W+PA and the fact that there is very little
literature to review on the other constructions, it seems reasonable to begin by reviewing

the literature on the prepositional phrase constructions.

2.2.1. NIW+PP

The focus of this section is on N/W+PP; more speciﬁcally, this section examines the
variety of PPs that may directly follow need/want, reviews the scant literature on these
constructions, and further elaborates the research questions that are addressed in the
present study. Need/want can occur with prepositional phrases beginning with down, in,

into, off on, over, through, to, under, up, among others, as in 25-34.

(25) Lisa needs/wants down the slide.
(26) Dean needs/wants in the building.
(27) He needs/wants into the club.
(28) Peter needs/wants off the plane.

(29) I need/want off this horse.

l4

(30)
(31)
(32)
(33)

(34)

The kids need/want over the creek.
Janet needs/wants through the crowd.
The baby needs/wants to John.

We need/want under the blanket, too.

Shelly needs/wants up the hill.

The construction N/W+PP seems to have escaped notice by language scholars and

even language pundits, for that matter. No dictionaries or grammar books that I consulted

have entries for N/W+PP, and only a few include examples of W+PP under entries for

W+PA (Spears 1987; K. Wilson 1993; OADLG 1999). Moreover, I have found no

publications of any kind that cite N+PP and less than a handful that cite W+PP (Pollard

1915:243; Burnham 1926:395; Krumpelmann 1939:156; cited in Wentworth 19442690;

K. Wilson 1993). The forms of W+PP found in the literature are headed by only four

prepositions: in, into, on, and out, as in 35-41.

(35)
(36)
(37)
(38)
(39)
(40)

(41)

‘I’m just a wantin’ in a higher grade’ (Pollard 1915:243).

‘I want in your class next semester’ (Krumpelmann 1939: 156).

‘wants in on the deal’ (OADLG 1999:1140)

‘Lardner wants into the Harding cabinet’ (Bumham 1926:395).

‘I want into that deal; it looks like a winner’ (K. Wilson 1993).
‘Everybody who wanted on the band-wagon . . .’ (Bumham 1926:395).

‘The children Want out of the house so they can play’ (Spears 1987:348).

15

Although N/W+PP has received little mention in the literature, the construction is
a regular part of my idiolect, and as I will show, not at all exotic in American English.
Furthermore, with little effort, examples of W+PP can befound in print, even in
nationally distributed periodicals. Two recent examples of want oﬂ'+ NP occurring in the
titles of articles in The Detroit News and Internetweek, located via a ProQuest search, are

seen in 42-43, respectively.

(42) ‘Want off the darn spam e-mail list?’ meRamus 2002).

(43) ‘IT wants off the Windows upgrade merry-go-round’ (Karpinksi 2002: 15).

The status of N/W+PP in the United States is unclear. This study will attempt to
gain a better understanding of the geographic, social, and stylistic distribution of
NIW+PP by addressing the following questions: Are forms of NIW+PP in common use in
the United States? Are these constructions more common in one or more geographic
regions? Are these constructions more commonly used by certain social groups or at
certain stylistic levels? Are some forms of N/W+PP more commonly used than others?
Any hypotheses made with respect to N/W+PP must, by necessity, rely on the literature

of N/W+PA or other studies of language variation.

2.2.2. NIW+PA
This section presents an overview of N/W+PA, by identifying the array of PAs that may
occur in these constructions, examining their coverage in the literature, and discussing the

conclusions that may be drawn ﬁ'om the treatment of N+PA as compared to W+PA in the

16

literature. Both need and want may be followed by a number of prepositional adverbs

including back, by, down, in, oﬁf on, out, over, through, under, up, and others as in 44-54.

(44) Is my old job still available? I need/want back.

(45) I’m sorry to bother you, but I need/want by.

(46) The baby is on the couch and needs/wants down.

(47) The cat needs/wants in.

(48) The fumes from this bus are making me sick. I need/want off.

(49) Hold the elevator. I need/want on.

(50) Dan was getting nervous in the stuck elevator. He needed/wanted out.
(51) Could you put a board over the creek? I need/want over.

(52) Can you make a path in the crowd? Mike needs/wants through.

(53) Try lifting the fence a little higher. I need/want under.

(54) Do you need/want up?

There is a clear bias in the literature towards the forms of W+PA over N+PA. Of more
than 60 sources that have cited one or more of these constructions, all have given want
with one or more prepositional adverbs, while only two have included need with a
prepositional adverb following. Not a single one of the many dialect words lists consulted
had an entry for N+PA. In only one of the dialectology projects is there any mention of
N+PA: In LAGS, two respondents (one from northeastern Georgia and one from
southeastern Georgia) gave need out in response to the prompt ‘The cat goes over to the

door and meows; you say, “The cat ”’ (LA GS concordance). Furthermore, no

17

usage guides that I consulted had entries for or examples of N+PA, with the notable
exception of DARE. DARE does have an entry for N+PA, in which the only citation is
one of the LAGS tokens (Cassidy & Hall 1996:769). Is it the case that N+PA occurs so
rarely as to not raise any interest? The present study seeks to explore this and other
questions with respect to the use of N+PA in different geographic regions, social groups,
and styles. In addition, the use of N+PA will be compared to W+PA. As was the case
with NIW+PP, any hypotheses made about the use of N+PA require deference to the
literature of W+PA and other studies of language variation.

Turning to W+PA, many publications have identiﬁed want as being able to take
one or more prepositional adverbs; however, in most sources, ﬁve prepositional adverbs
are cited most frequently. By far, the most commonly cited forms are want in and want
out. More than 40 usage guides, dialect word lists, and dialectology projects include
citations of want in or want out, and nearly all of these mention both (Carruth 1892:142;
Rice 1902:248; Crumb 1903:336; Carr 1904:422; Pound 1905:66; Prettyman 1907:76;
Heydrick 1908:51; Harvey 1914:165; Pollard 1915:243; Lehman 1921:110; Mullen
1925:149, 152; Burnham 1926:395; Jacobson 1931:19; Maxﬁeld 1931:20; Stanley
1936b:353; Aurand 1939:30; Krumpelmann 1939:156; McAtee 1942:70; Warnick
1942:16; Griebling 1947:5; A. Wilson 1948:238; Ashcom 1953:255; McAtee 1956:49;
Evans & Evans 1957:545; Webster’s, 3rd ed., s.v. ‘want’; Allen 1975:72; Shaw 1975:252;
McDavid & Payne 1976; Morris & Morris 1985:614; Quirk et al 1985:732; Random
House 2'“1 ed., s.v. ‘want in or out’; Spears 1987:348; Greenbaum & Whitcut 19882762;
OED 2"d ed., s.v. ‘want’; K. Wilson 1993). In addition, want up and want down are often

cited, particularly in dialect word lists (Carr 1904:422; Pound 1905:66; Prettyman

18

1907:76; Heydrick 1908:51; Stanley 1936bz353; Aurand 1939:30; McAtee 1942:70;
Warnick 1942:16; Griebling l947:5; McDavid & Payne 1976; K. Wilson 1993). Thanks
to its inclusion in the worksheets of the Linguistic Atlas projects and DARE, want of has
received a fair amount of attention not only in publications using the records from these
projects (e. g. Shuy 1962; Allen 1975) but also in studies that emulated the methodology
and used the worksheets of the Linguistic Atlas projects (e.g. Hartman 1966; Foley 1972).
Moreover, want off is very nearly the only form to be systematically sampled at
all let alone in a variety of regions. Only two other forms of W+PA, want in and want
out, have been sampled, albeit meagerly. At the discretion of individual ﬁeldworkers,
want in and want out were elicited in LANCS and LAUM (Allen 197 5:71), as was want
out in LAGS (Pederson 1991). A number of prepositional adverbs received mention in a
smaller number of sources including want on (Pound 1905:66; Stanley1936b2353;
McAtee 1942:70; Allen 1975:71-72; Dakin 1966bz513), want back (Burnham 1926:395;
McAtee 1942:70), want by (Burnham 1926:395), want abroad (Wilson 1853 cited in
Eliason 1956:145), want over (Carlton 1843:91), and want through (Griebling l947:5).
As seen here, previous literature provides a wealth of information on W+PA but,
at the same time, is severely limited in scope. The literature has primarily focused on
W+PA, to the exclusion of N+PA and NIW+PP and on a small number of prepositional
adverbs. In addition, even fewer prepositional adverbs have been systematically studied.
Several questions about the use of need and want with various prepositional

adverbs stems from the literature (or lack of it):

19

(1) Is W+PA more acceptable within a particular region and geographically more
widespread than N+PA?

(2) Can the geographic, social, and stylistic distribution of one of these forms, e. g.
want oﬁf be taken as representative for all forms of NIW+PA?

(3) Do the prepositional adverbs have different geographic, social, and stylistic
distributions?

(4) Can we assume that the most written about forms of N/W+PA, namely want in,

want out, and want of are more common and more widely used than other forms?

The sheer volume of literature alone may lead to some hypotheses. For example,
the fact that there are myriad citations of W+PA in the literature and very few of N+PA
and NIW+PP may permit the hypothesis that W+PA is more acceptable than N+PA and
NW+PP; that is, if it is assumed that more citations and attention to a construction
translate to greater use. Furthermore, the fact that the majority of the literature focuses on
want with a few prepositional adverbs may suggest that these forms are the most
widespread. A few sources even state that some prepositional adverbs are more common
than others. For example, in the Harper Dictionary of Contemporary Usage (Morris &
Morris 1985:614) want out is described as ‘a regional expression heard most commonly
in the Middle West’ while want in is ‘a similar phrase, if less common.’ In LA UM, Allen
(1975:71) stated that ‘want in is even more frequent than want oﬂ’ in the Midland. Based
on these comments, one might hypothesize that in the Midland, a usage hierarchy exists

whereby want out is most common, followed by want in, then want of. The present

20

study seeks to uncover any such distributional differences and hierarchies among N+PA

and W+PA as well as various prepositional adverbs used with need/want.

2.3. Geographic distribution

This section examines the geographic distribution of W+PA using data primarily from
dialectology projects and dialect words lists. In this section and throughout the
dissertation, I refer to three major dialect areas in the United States, the North, the
Midland (North Midland and South Midland), and the South, as deﬁned in Figure 2.1
(McCrum et al 1986:23 8). Over the past several years, debates about the status and
etiology of the dialect regions in the United States have continued (see, e. g. Murray &
Simon 1999:153-155; Benson to appear), and dialect maps of the US. have been drawn
and redrawn based on different linguistic features; for dialect maps based primarily on
lexical data, see Carver 1987; for dialect maps based primarily on phonological data, see
Labov et a1 1997, to appear. One map that incorporates phonological, lexical, and
grammatical data from the Linguistic Atlas projects into its drawing of dialect boundaries
in the United States (without too many subregions obscuring the major dialect regions, as
in Lance 1994) is shown in Figure 2.1. The use of these dialect boundaries not only
allows for reﬂection of their current validity but also facilitates comparisons to other
dialect studies that have also adopted them (e.g. Murray 1993; Frazer et a1 1996; Murray

& Simon 1999, 2002).

21

 

North

 

 

 

N. Midland

 

 

 

S. Midland

 

 

South

 

Figure 2.1. US. dialect areas (Mch et a1 1986:238)

The discussion of the geographic distribution relies almost eXclusively on the
literature of W+PA, since very little exists for NIW+PP and N+PA. In spite of the fact
that a substantial number of publications include citations of W+PA, there is no generally
accepted understanding of its geographic distribution. A look into several usage guides
reveals that some associate W+PA with a regional distribution, most commonly Midland
(e.g. Evans & Evans 1957:545; Morris & Morris 1985:614; Quirk et al 1985:732;
Random House 2'“I ed., s.v. ‘want in or out’; K. Wilson 1993): ‘1 want out instead of I
want to get out is a regional expression heard most commonly in the Middle West’
(Morris & Morris 1985:614). Other usage guides state directly (as in the OED) or imply,
by not giving any geographic descriptors, that the construction is in common use across

the United States (e.g. Marckwardt & Walcott 1938:49; Webster 's, 3rd ed., s.v. ‘want’;

22

Shaw 1975:252; Spears 1987:348; Greenbaum & Whitcut 1988:762; OED 2“d ed., s.v.
‘want’; American Heritage, 4th ed., s.v. ‘want out/in’). In the ﬁrst and second editions of
the OED, want out was described as ‘U.S. colloquial’. More than 70 years after the ﬁrst
such proclamation and many more conflicting statements later, the question of whether
W+PA is more frequently used in one particular dialect area or is in widespread use

remains unanswered.

2.3.1. The Midland

Data from dialectology projects and dialect word lists show robust use of W+PA in the
Midland. Linguistic Atlas project data and DARE data present strong evidence of a
Midland distribution for W+PA. The only form to be systematically elicited, due to its
inclusion in the questionnaires of both the Linguistic Atlas projects and DARE, was want
oﬂ.’ The prompts for the elicitation of want of are similar in the Linguistic Atlas projects
and DARE: ‘You tell the bus driver, “The next corner is where I want ”’
(Pederson 1974: 1 89); and ‘What you’d say to a bus driver: “Please stop at the next
comer—I want .”’ (Cassidy 1985:1xxix). Some respondents gave want out as a
response to these prompts, as well. In addition, other prompts were used at the discretion
of ﬁeldworkers to elicit want out and/or want in in LA UM, LANCS, and LAGS, e.g. ‘The
cat goes over to the door and meows; you say, “The cat ”’ (Pederson et al
1974:189). Unfortunately, the data on want out/in are limited since few ﬁeldworkers
actually employed the additional prompts, and it is not clear which respondents were

chosen for additional prompts. On occasion, incidental uses of W+PA were also

recorded. Figure 2.2, displays the attested uses of want oﬁf want out, and want in from

23

the Linguistic Atlas projects in the central and eastern United States, i.e., LANCS,
LAMSAS, LANE, LA UM, and LAGS (see Appendix C). As shown in Figure 2.2, the use of
want oﬁ’, want out, and want in was concentrated in the Midland (Kurath 1943:map 699;
Kurath 1949:map 159; Allen 1975271; McDavid & Payne 1976; LAGS concordance).

Linguistic Atlas data from much of the area west of the Mississippi River are
difﬁcult to ﬁnd and access. Nevertheless, a few statements about the use of want of in
the West may be drawn from the Linguistic Atlas of the Paciﬁc Northwest and studies
using the Linguistic Atlas project data. In the Paciﬁc Northwest (Washington and Idaho),
want 017 was used by nearly 50% of respondents in Washington; the majority of want of
users were clustered in the northwest part of the States around Seattle and Tacoma. Only
two respondents in Idaho used want of. In Colorado, want of was described as
‘infrequent’ overall and more characteristic of the east-central parts of the state (Hankey
1960211, 41), and in California and Nevada, want of was used by 15% of respondents
and Was characterized as scattered (Bright 19712201).

DARE data also show greater use of want of and want out in the Midland in states
east of the Mississippi River than in the North, the South, and states west of the

Mississippi River (Hall, p.c.), as shown in Figure 2.3.

24

6 522:? use 9% Hosea 83. ceases: a «dis .3 3am

5 Em? 0
So :33 4
be :83 O .

 

 

 

 

25

In other sources, myriad observations of W+PA in the Midland can be found. The
earliest literary citations of W+PA even identify Midland areas as the home of the
construction. These mid-nineteenth century citations are not simply causal or unaware
uses by the authors but are, in fact, quoted speech attributed to long time residents of
Indiana and Illinois. The New Purchase, a dramatized autobiography set in Indiana
(Bloomington area), has two such forms: ‘she “know’d we wanted over’” attributed to a
local ferrywoman by the narrator, a recent settler to the area, and "’Maybe,” she says to

9”

herself, “its some poor Injin wants in attributed to an early settler in a story about her
encounter with a bear (Carlton 1843 :91, 170). In fact, in a footnote added in the 3rd
edition of The New Purchase, editor James A. Woodburn explained that, “‘I want over,”
“I want in,” “I want out”, etc., are pioneer forms of speech that are still not uncommon in
certain regions of the Middle West’ (Carlton 1916:78). Another early citation of W+PA is
found in Zury: The Meanest Man in Spring County, set in Illinois (see also Frazer 1982).
In an exchange between an unruly schoolboy and a teacher from the East, want out is
explicitly identiﬁed as a regional form:

“‘Please ‘m’, m’ I g’aout?”

“What did you say?”

“I want ou’doors.”

“You want outdoors.”
“That’s wut I said! I want aout.

‘At last she understood this [Middle-Westernism], new to her. To “want out” is to desire
to go out’ (Kirkland 18872157). Additional examples of W+PA can be found in countless
literary works and periodicals (e.g. citations in Wentworth 1944:689-690; OED 2nd ed.,
s.v. ‘want’; Ashcom 1953:255; Cassidy & Hall to appear). Although literary uses may not

be considered actual uses, they are often good examples of stereotyped linguistic features.

27

The use of such linguistic features in literature, especially when there is overt
commentary on the use, demonstrates an acute level of awareness, in this case of W+PA
as a regional/Midland form as early as the late 18003. Additional evidence for a Midland
distribution of W+PA may be found in nonliterary works, as well.

Dialect word lists and independent studies (i.e. those not using dialectology
project data) also contained abundant reports of W+PA in the Midland. As shown in
Table 2.1, the greatest number of observations of W+PA in a given area came from
Pennsylvania, where want + down/in/out/up were often directly associated with
Pennsylvania German communities. Observations of W+PA were also given in other
Midland areas including the states of Illinois, Indiana, Kansas, Maryland, Missouri,

Nebraska, Ohio, Tennessee, and West Virginia.

Table 2.1. Dialect word list reports of W+PA in the Midland

 

 

 

 

 

 

 

 

 

 

 

 

States back by down in of on out through up' References
Illinois X X Rice 19022248
Indiana X X X X X McAtee 1942270
Kansas X X X X Carruth 18922142
Burnham 1926:395
Maryland X X X Warnick 1942:1
Missouri X X Crumb 1903:336
Nebraska X X X X X X Pound 1905:66
Ohio X X X X X X Griebling l947:5
Hartman 1966:72
Pennsylvania X X X X Prettyman 1907:76
Heydrick 1908:5 1
Maxﬁeld 1931:20
Aurand 1939230
A.Wilson 19482238
Ashcom 1953:255
Tennessee X Pollard 19152243
W. Virginia X X X X Krumpelmann
1939:156
cited in Wentworth
1944:689-690

 

 

 

 

 

 

 

 

 

 

 

28

 

Taken altogether, dialectology project data, dialect word lists, and literary sources
provide strong evidence for the use of various forms of W+PA, in particular want in, oﬁf

and out, in the Midland.

2.3.2. The South

There is little evidence of W+PA in the South (Kurath 1943:map 699; Kurath 1949:map
159; Allen 1975271; McDavid & Payne 1976; LAGS concordance; Hall, p.c.). As shown
in Figure 2.2 and Figure 2.3, in both Linguistic Atlas and DARE data, there are no
attestations of want oﬂ and want out in eastern North Carolina and eastern South Carolina
and only a smattering in western South Carolina, Georgia, Florida, Alabama, Mississippi,
Louisiana, and east Texas, with the heaviest concentration in northern Alabama and east
Texas. A few dialect words lists and other studies reported on the use of W+PA in

Alabama and east Texas, as shown in Table 2.2.

Table 2.2. Dialect word list reports of W+PA in the South

 

 

 

 

 

States down in of on out References

Alabama X Foley 1972223

Texas (east) X X X X X Stanley l936b:353
Norman 19712140

 

 

 

 

 

 

 

Based on the data from dialectology projects and dialect word lists, in spite of a few
pockets of greater use in the South, I come to the same conclusion as Kurath (1949230)
that the use of W+PA is not at all strong in the South.

In spite of the fact that the use of W+PA is not strong in the South, the South is

the only area to have an observed use of N+PA. Need out was given as a response to the

29

prompt: ‘The cat goes over to the door and meows; you say, “The cat ”’
by two respondents, one in northeast Georgia and the other in southeast Georgia (LA GS

concordance).

2.3.3. The North
According to dialectology project data and dialect word lists, W+PA was virtually
unknown in the North. In the Linguistic Atlas project data and the DARE data, shown in
Figure 2.2 and Figure 2.3, the nearly blank North does not mean that the states in the
North were not sampled; on the contrary, they were richly sampled in LANE, LANCS,
LA UM, and DARE. In LANE, there were no occurrences of want oﬂ in the northeastern
United States, though there was one in New Brunswick, Canada (Kurath 1943:map 699).
In LANCS, only two respondents in Michigan and none in Wisconsin used want of
(McDavid & Payne 1976). In LA UM, one respondent in southern Minnesota used want
of and a handful of respondents in South Dakota used either want oﬂ,’ want in, or want
out (Allen 1975:71). This greater use of W+PA in South Dakota is surprising not only
because the use there is stronger than anywhere in the North but also because it seems to
run counter to the trend in the Midland of decreasing usage from east to west.
Furthermore, there were almost no observations of W+PA in dialect word lists or
other studies. Only one observation was found in a dialect word list. In Milwaukee, want
in/out was said to be characteristic of the local dialect (Jacobson 1931:18-19). Overall,
previous research gives little indication that W+PA is used or even known in the North.

Kurath’s (1949230) statement about the distribution of want of seems validated by the

30

literature: ‘The records of the Linguistic Atlas show that I want oﬁr is in common use in
nearly all of the Midland [. . .], and that it is not current at all in the North and the South.’

The lack of attestations of W+PA in the North and the South calls into question
depictions of the W+PA, like in the OED ’s label ‘colloquial U.S.’, that imply
widespread use. Additional research is needed to shed light on the geographic distribution
of W+PA and related constructions in the United States.

Dialectology projects and studies as well as dialect word lists and reports provide
strong evidence for a primarily Midland distribution of W+PA and by extension N+PA
and N/W+PP. Additional support for a Midland distribution of these constructions may
be found in other grammatical constructions with a principally Midland distribution. The
behavior of NIW+PP and N/W+PA may well parallel that of need/want/like + past
participle, as in 55-57, and positive anymore, as in 58, all of which show robust usage in
the Midland (e.g. Youmans 1986:61; Murray 1993:177-185; Frazer et al 1996:260-265;

Murray & Simon 19992144-150, 2002:37-44).

(55) The house needs painted.
(56) The kids want fed.
(57) The dog likes petted.

(58) We go to the movies a lot anymore. (53)
The literature on these constructions provides evidence of not only similar geographic

distributions but also similar origins. The presence of need/want/like + past participle,

positive anymore, and W+PA in the United States have been reliably traced to the

31

Scotch-Irish (Stanley 1936a23-4; Marckwardt 194829; Crozier 1984:326; Montgomery
1991, 19972196; Murray 19932185; Adams 2000:300; Murray & Simon 2002:45-53).

Another interesting parallel with need/want/like + past participle (see Murray &
Simon 2002) is that the presence of W+PA in the United States has at times been
attributed to German, speciﬁcally Pennsylvania German, inﬂuence (e.g. Krapp 1925:256;
Krumpelmann 1939:156; Griebling 194725; A. Wilson 1948:238; Adams 2000:300),
which is not surprising because of the overlapping settlement patterns, to some extent, of
the Germans and Scotch-Irish. Nevertheless, the Scotch-Irish inﬂuence on the language
of the area has dominated any German inﬂuence, and linguistic features that had been
assumed to be of German origin have been argued using data from dictionaries, personal
diaries and correspondence, and other sources to be of Scotch-Irish origin (Montgomery
1991, 1997; Murray & Simon 2002). As a result of their Scotch-Irish origins and their
similar geographic distributions in the United States, one might expect other parallels
between NIW+PP and N/W+PA, on the one hand, and need/want/like + past participle
and positive anymore, on the other.

In closing, my hypothesis with respect to the geographic distribution of W+PA
and the related constructions of N/W+PP and N+PA in this study is that these

constructions have considerable Midland distributions.

2.3.4. Geographic expansion
The paucity of attestations and observations of want in/out in the North, and to some
extent the South, is at odds with numerous publications that have claimed widespread

distribution for these constructions (see Section 2.3). Perhaps some of the confusion

32

about the geographic distribution of W+PA is the result of a change in progress. The data
from dialectology projects, dialect word lists, and other studies provide ﬁrm support for a
Midland distribution of W+PA, at least historically. Not all Midland respondents of the
Linguistic Atlas projects or DARE used want off, which may be due, in part, to the
methodology, the skill of the researcher, or some other factor. In any case, it appears that
there are nonusers in the Midland to whom the construction could still spread. In southern
Ohio, Hartman (1966272) observed that ‘regional “want off” is pushing out the standard
expression.’

Over the past 70 years, a number of publications have claimed a more widespread
distribution for W+PA. The ﬁrst edition of the OED described want in/out as ‘U.S.
colloquial’, and based on that, Marckwardt & Walcott (1938249) in response to Leonard’s
(19322124) description of The kitten mews whenever it wants in as illiterate, categorized
want in as ‘American Colloquial Eng’. Speciﬁcally relating to the spread of W+PA to the
North, Griebling (l947:5) stated that in the early twentieth century, want in was common
in Columbus, a linguistically Midland city, and rare in Cleveland, 3 linguistically
Northern city, but by mid-century, want in was in widespread use in Cleveland. Ashcom
(1953:255) stated that ‘the spread of the construction [W+PA] is so wide that it can
scarcely be considered regional any longer’. Hankey (1961:269) said that want of may be
‘becom[ing] more frequent and widespread’. The claims of a widespread distribution for
W+PA have become more numerous in recent years (e.g. Webster ’s, 3rd ed., s.v. ‘want’
1981; Shaw 19752252; Quirk et al 1985:732; Spears 1987:348; Greenbaum & Whitcut

19882762; OADLG 1999:1140).

33

Current print media sources provide additional evidence for a wide-reaching
distribution of W+PA. In the print media, instances of want out and want in, at least, are
not at all uncommon. In 59-62, I cite a few recent occurrences, found using ProQuest, in
the titles of articles appearing in nationally and internationally distributed periodicals, a

context which would seem to come under the close scrutiny of editors:

(59) ‘When parents want out’ (The Unesco Courier: Noguera 2000224).
(60) ‘Kodak wants out of the classroom’ (Information Week: Goodridge 2001 :94).
(61) ‘Serving the wealthy: Everybody wants in’ (ABA Banking J: Asher 2001242).

(62) ‘Why Libya wants in’ (Time: Zagorin 2000:66).

These sources strongly suggest that the use of W+PA may be spreading beyond
the Midland. It is not without precedent that Midland forms, particularly grammatical
forms, have spread to other dialect areas. Recent research ﬁndings on need/want + past
participle (as illustrated in 55-56) and positive anymore (as illustrated in 58),
grammatical structures with similar origins that are well-rooted in the Midland, indicate
that the use of these constructions is indeed increasing within and moving beyond the
Midland (Labov 1973272; Youmans 1986274; Labov 19912277; Murray 19932175, 185;
Frazer et a1 19962268; Murray & Simon 2002257). In addition, previous research on the
interaction between social factors and W+PA discussed in section 2.5 may lend more

support to an argument for change in progress.

34

2.4. Semantic features

An important semantic distinction, namely concrete versus abstract, may play a role in

the geographic distribution of these forms. To better understand the issue, compare

sentences 63-68 with sentences 69-74. Sentences 63-68 all illustrate concrete uses:

Physical movement is intended or required by the subject of need/want, and the location

identiﬁed by the prepositional phrase/prepositional adverb is a concrete, physical space.

(63)
(64)
(65)
(66)
(67)

(68)

Dean needs in the building. (1)

Do you want off the merry-go-round? (14)

The cat needs in. (21)

The kids say it’s too hot outside, and they want in. (3 7)

When the roller coaster picked up speed, Ed yelled that he wanted off. (41)

Mike needs out at the comer. (30)

Sentences 69-74 depict abstract uses: No physical movement is involved; rather a desire

to be involved or no longer involved is the intent of the subject of need/want, and the

location identiﬁed by the prepositional phrase/prepositional adverb is abstract.

(69)
(70)
(71)
(72)

(73)

Katy needs in this class in order to graduate this semester. (4)

Sam wanted in the Army. (12)

The new drug experiment is Monica’s last hope. She needs in. (24)
That sounds like a great plan. I want in. (40)

The baseball team hasn’t won a game in several weeks, and Jim wants off. (43)

35

(74) Barry’s job has become too stressful; he needs out. (31)

Interestingly, the concrete/abstract distinction is hinted at in some dictionaries
under entries for want in/out, most notably Random House, American Heritage, Harper
Dictionary of Contemporary Usage, and American Idioms Dictionary. One of the most
lucid descriptions is found in the Harper Dictionary, ‘ [want out] is used literally in terms
of physical departure but also ﬁguratively in the sense of “to be free of involvement’”
(Morris & Morris 1985:614). None of these sources, however, identify distributional
differences for concrete and abstract senses of W+PA. Random House and Harper
Dictionary describe both uses as Midland; American Heritage states that both are
informal, while the American Idioms Dictionary puts no qualiﬁcations on their use.

Note that all examples of W+PA that were found in the titles of nationally
distributed periodicals (presented in 59-62) are indeed abstract senses. In fact, all of the
more recent examples that are cited in the OED (2"d ed., s.v. ‘want’), the most well-
known, most respected, and earliest source to claim widespread use for want out, are in

the category of abstract senses, as shown in 75-78:

(75) ‘One of the kids who had paid his money wanted out’ (1973).

(76) ‘Britain may just be weary of industrial growth and may be saying in quite a
sophisticated way that it wants out regardless of the cost’ (1973).

(77) ‘In recent weeks the Federal Reserve chairman, Mr. Paul Volcker, has reportedly
told friends that he wants out and would be interested in the presidency of the
World Bank’ (1984).

(78) ‘If you want out, it is possible to live, if only internally, a better life’ (1985).

36

Three points, in particular, lead me to believe that concrete senses and abstract
senses of W+PA and related constructions may have different geographic distributions in
the United States and may, in fact, be partly responsible for the lack of uniformity in
descriptions of the distribution. First, the distinction between concrete and abstract has
been shown to play a role in semantic change, in what is often called semantic or
metaphoric extension (e.g. Williams 19752153-211; Traugott 1982, 1995; Langacker
1999; Liidtke 1999). Second, previous research suggests that W+PA is commonly used in
the Midland and rarely used in the North and the South. Third, the superregional
examples of W+PA that I have encountered share the characteristic of an abstract ‘goal’
or location identiﬁed by the preposition. As previously noted, the examples of W+PA
found in nationally distributed sources are all abstract senses. The uses of want oﬂ, want
out, and want in from the Linguistic Atlas projects and DARE (outlined in Section 2.3)
are all concrete senses. Of the examples given in dialect word lists, most are also
concrete. Taking all of this information together, I hypothesize that concrete senses of
W+PA and related constructions show more regional, i.e. Midland, distribution, while
abstract senses show more widespread distribution. Additionally, I hypothesize that
within each dialect region, and particularly in the North and the South, the use of abstract
senses of N/W+PP and N/W+PA will be greater than that of concrete senses. The
hypothesis appears to run counter to the general tendency in semantic change for
linguistic items to move from concrete to abstract: ‘Words for abstractions will generally
develop out of words for physical experience’ (Williams 19752207). This issue is

explored in more detail in Section 4.2.

37

2.5. Social factors

In this section, the focus is on social factors as independent variables and their
relationship to linguistic variables. For well over 70 years, linguists have made
connections between the social characteristics of speakers and language use. In the
Linguistic Atlas projects, respondents were classiﬁed into one of three ‘types’ based on a
combination of social factors, such as age, level of education, social contacts (e.g. Kurath
1939244; Allen 197523). More recently, social variables such as age, level of education,
gender, socioeconomic status, network ties and others have become important
components of studies of language variation. In this study, four social variables are
considered: rural-urban residence, gender, socioeconomic status, and age. This section
outlines the general, theoretical concerns for each variable and reviews the existing

literature on their relation to the distribution of W+PA.

2.5.1. Rural-urban residence

It has long been said that dialect differences at all levels of the grammar exist between
rural and urban areas (Wolfram & Schilling-Estes 1998230). For example, in a study of
postvocalic [r] in South Carolina, rural respondents were found to be more [r]-ful
constriction than urban respondents, who tended to be more [r]-less (McDavid 19642476).
Furthermore, the rural-urban distinction has often played an integral role in analyses of
language change: urban areas are typically, but not always, centers of change (Bailey et
a1 1993:384-385; Wolfram & Schilling-Estes 1998:30; Labov 20012437). Two examples
of the rural-urban distinction in language change are evident in sound changes currently

in progress in the United States. The Northern Cities Chain Shift, as can been seen from

38

the name, has its locus in urban areas and only recently has begun spreading to less urban
areas (e. g. Eckert 19912219; Labov et al 1997; Labov 20012285; Labov et al to appear).
Though much less has been written about it, the Southern Vowel Shift, on the other hand,
is centered in less urban areas (F ridland 1998; Labov et al to appear). The particulars of
these vowels shifts are not important to the present study; what is important is that these
vowel shifts illustrate the role that the rural-urban distinction continues to play in
language variation and change.

Only a few studies of W+PA collected rural-urban demographic information
about the respondents. DARE data show a continuum of increased use of want of from

urban to rural areas, as presented in Figure 2.4 (Hall, p.c.).

30% 1
25% i 25.1%
20% -

15%

10% r

 

5%:

 

 

0% t . , T
Urban (N=64) Large City (N=52) Small City (N=201) Village (N=39l) Rural (N=279)

Figure 2.4. Use of want oﬂ in DARE by community type (Hall, p.c.)

39

More urban areas have a lower rate of use of want oﬂ, while more rural areas have
a higher rate of use. A study of the speech in HOcking County in southern Ohio, an area
with dense want oﬂ use, contradicts the DARE ﬁndings, which were for the entire United
States. Hartman (1966:72) found slightly greater use of want of among urban
respondents as compared to rural respondents: 75% (N=4) of urban respondents and 63%
(N =16) of rural respondents used want oﬁ’. The difference between the DARE ﬁndings
and Hartman’s ﬁndings may have to do with the fact that the DARE ﬁndings are based on
statistics from the entire United States, including areas where want of has not been
shown to be in widespread use, while Hartman’s ﬁndings are from one county, situated in
an area where want oﬂ has been shown to be common.

Additional support for hypothesizing a lack of a rural-urban distinction in the use
of W+PA may be found in the distribution of other grammatical variables that are also
common in the Midland. Studies of need/want/like + past participle (as seen in 55-5 7),
have found that these constructions are used at nearly the same rate in rural and urban
areas alike (Frazer et al 19962265-266; Murray & Simon 19992150, 2002253). The
ﬁndings of need/want/like + past participle primarily come from states in the Midland
and Inland North and were based on a combined-methodologies approach that minimally
included collecting data via surveys of university students, questionnaires by
telephone/mail, and postings on electronic bulletin boards (see Simon & Murray 1999).
In each case, the conclusion was the same: ‘Our data repeatedly show that the
construction is not restricted to rural or outlying areas but that it regulme appears in

larger cities such as Indianapolis, Kansas City, Salt Lake City, St. Louis, Columbus,

40

Pittsburgh, and Philadelphia’ (Frazer et al 19962265). My hypothesis is that W+PA and

related constructions are not distributed differently in rural and urban areas.

2.5.2. Social class

The focus of this section is on the interrelationship between social class and language
variation. For nearly 40 years, socioeconomic status as a social variable has been shown
time and time again to play an important role in the study of language variation and
change (e.g. Labov 1966273-76; Trudgill 1974:97-99). If a linguistic feature is evaluated
positively or negatively (above or below the level of consciousness), it is expected that
that feature will also show some type of social stratiﬁcation in its use. Two general types
of stratiﬁcation among social classes have emerged: gradient stratiﬁcation and sharp
stratiﬁcation. Gradient stratiﬁcation shows a gradual increase or decrease in the use of a
feature from one social class to the next and is more common for phonological variables.
Sharp stratiﬁcation, on the other hand, shows a jump or drop in the use of a feature
between the working and middle classes, and is more common for grammatical variables
(Wolfram & Schilling-Estes 19982155-157).

The dialectology projects are among the few studies that have been concerned
with the distribution of want + prepositional adverb across socioeconomic classes. Early
reports claimed that the use of want oﬁ' varied based on social class/level of education.
More recent studies, however, have found little or no differences in use based on social
class. Griebling (194725) was among the ﬁrst to claim that W+PA may be socially
diagnostic in his statement that this construction is ‘indeed very common among the

relatively uneducated people of Ohio and Pennsylvania.’ Early Linguistic Atlas

41

publications concluded that want of?r was commonly used except by the well-educated
and ‘cultured’ (Kurath 1949230; McDavid 19582519). In subsequent publications using
data from LAUM and LANCS, Allen (1975272) and Dakin (l966b2515) refuted this,
noting that they could not detect any differences in use by social class.

Allen’s and Dakin’s ﬁndings are paralleled by recent research on other
grammatical variables with strong Midland distributions. Positive anymore (illustrated in
58), one of the most studied features of grammatical variation in the United States, has
little or no socioeconomic stratiﬁcation in regions where it is accepted (Labov 19912277;
Murray 19932185). Recent work on need/want/like + past participle (illustrated in 55-57)
has shown that these constructions are also not socioeconomically diagnostic (Frazer et al
1996:264-265; Murray & Simon 1999:149-150, 2002253, 55). As illustrated in Table 2.3,
surveys of respondents primarily from Midland and Inland North states found that the use
of the constructions was nearly ﬂat across the social classes: 41% for need + past
participle, 22% for want + past participle, and 5% for like + past participle (Frazer et a1
19962265; Murray & Simon 19992150, 2002255).

Table 2.3. Use of need/want/like + past participle by socioeconomic class (Frazer et al
19962265; Murray & Simon 19992150, 2002255)

 

 

 

 

 

31:1: rsoeconormc need + past participle want + past participle like + past participle
Upper 40.8% (863) 21.6% (60) 4.7% (12)
Middle 41.0% (2,810) 22.3% (77) 5.3% (17)
Lower 41.3% (1,342) 21.7% (63) 5.0% (13)

 

 

 

 

42

 

Grammatical constructions like positive anymore and need/want/like + past
participle are unique in their ﬂat socioeconomic distributions within the Midland, while
at the same time evoking negative reactions by nonusers within and outside of the
Midland (Wolfram and Christian 19762105; Murray 19932183; Frazer et al 19962266;
Murray & Simon 19992150-151, 2002253). Based on these ﬁndings, my hypothesis is
that W+PA and related constructions also show no signiﬁcant differences in use across

socioeconomic classes.

2.5.3. Age
In this section, the role of age in the study of language variation is presented. The study
of language variation across different age groups has played an integral role in
determining the status of a linguistic feature as stable or undergoing change. It is by now
well accepted that apparent time studies, i.e. studies of respondents of different ages at
the same point in time, are as reliable as real time studies of language change (e.g.
Trudgill 2002254, 61; Bailey et al 19912260, 263). When age variation of a linguistic
feature is found in an apparent time study, however, the possibility of age-grading, use of
different linguistic features at various life stages, must also be ruled out. Generally,
however, in studies of language change, the behavior of younger speakers is expected to
contrast with older speakers, who maintain their use of the conservative form and show
lesser use of the innovative form.

Almost nothing is known about the use of W+PA across different age groups. In
his study of the speech in southeastern Ohio, Hartman (1966272) found that want oﬂ was

used more by younger respondents (N =6), 83%, than by middle (N=8), 63% and older

43

respondents (N =6), 50%. From an apparent time perspective, these results suggest that
the use of want of may be increasing in the area, reﬂecting a change in progress. At the
same time, it must be recognized that the differences in use may be attributable to
Hartman’s small sample size. Furthermore, evidence from other Midland grammatical
constructions do not support age stratiﬁcation for W+PA and related constructions.

Positive anymore and need/want/like + past participle (as seen in 55-5 8) are not
stratiﬁed by age (Wolfram and Christian 19762105; Youmans 1986274; Murray 19932185;
Frazer et al 1996:264-265; Murray & Simon 1999:149-150, 2002253, 55). As shown in
Table 2.4, the use of need/want/like + past participle by age group shows little variation
(Frazer et al 19962265; Murray & Simon 19992150, 2002255).

Table 2.4. Use of need/want/like + past participle by age (Frazer et a1 19962265; Murray
& Simon 19992150, 2002255)

 

 

 

 

 

 

 

 

Age need + past participle want + past participle like + past participle
>20 years old 40.2% (1,621) 21.8% (49) 5.4% (11)
20-40 years old 41.4% (1,543) 21 .5% (55) 5.1% ( 12)
41-60 years old 41.8% (1,144) 21.5% (51) 4.5% (10)
61-80 years old 40.9% (704) 23.0% (45) 5.0% (9)

 

 

Based on these ﬁndings, my hypothesis of the use of W+PA and related constructions is

that they do not show age variation.

2.5.4. Gender
In this section, the role of gender in sociolinguistic studies of language variation is
explored. I recognize that deﬁning gender based on biological sex, as is the practice in

the ﬁeld of quantitative sociolinguistics, and even the use of the terms ‘sex’ and ‘gender’,

44

 

have been topics of much debate in recent years (e.g. Eckert 1989; Labov 1990; Eckert &
McConnell-Ginet 1992; Wodak & Benke 1997). These issues notwithstanding, the
purpose of this section is to outline general trends found to exist for gender as a social
variable in studies of language variation and to examine the role of gender and W+PA.
Two tendencies have emerged concerning the use of language features by men and
women. One generalization is that women tend to use more standard features than men,
while men tend to use more nonstandard features (e.g. Labov 1966:312-313; Chambers &
Trudgill 1980:72-73; Eisikovits 1989237-41; see Labov 20012261-293). The second
tendency concerns language change: women tend to more readily adopt new prestige
language features than men (see Labov 20012261-293). Both tendencies may potentially
be applicable to the current study. In order to make a hypothesis about a language feature
based on the ﬁrst tendency, an understanding of the social distribution or at the very least
the social position of the feature must be known. In order to make a hypothesis about a
language feature based on the second tendency, it must be known whether that feature is
in the process of change either through real-time or apparent time study. Previous
research on want of and the related constructions of need/want/like + past participle
(illustrated in 55-57) shows differing results, as shown in Table 2.5. In DARE data of
want oﬁf gathered from the entire United States, women showed greater use of want off
than men (Hall, p.c.). Studies of primarily Midland and Inland North areas of the use of
need/want/Iike + past participle found no gender differences in their use (Frazer et al

19962265; Murray & Simon 19992150, 2002255).

45

Table 2.5. Use of want of and need/want/like + past participle by gender (Hall, p.c.;
Frazer et a1 19962265; Murray & Simon 19992150, 2002255)

 

 

 

 

 

 

 

 

need + past want + past like + past
Gender want of participle . participle participle
Men 16.0% (70) 41.9% (2,478) 21 .7% (98) 5.2% (21)
Women 21.1% (116) 40.3% (2,541) 22.1% (102) 4.8% (21)

 

With respect to the use of N/W+PP and N/W+PA by gender, I hypothesize that gender

behaves like other social variables in that no variation in use is evident.

2.6. Style

Deﬁning ‘style’ is no easy task (e.g. Irvine 2001; Haynes 1995). While the phenomenon
that is of interest in this study comes under the umbrella of style (using any number of
current deﬁnitions), it is perhaps more appropriately called register: language variation
based on situations of use. Style/register has often been an important independent
variable in studies of phonological variation, in particular, phonological change (e. g.
Labov 19662396-399) but has less often been included in studies of grammatical variation
and change.

No one has actually examined the use of W+PA across different styles;
nevertheless, several publications, mainly usage guides, make claims about the
appropriate stylistic level for its use. The general consensus is that forms of W+PA,
speciﬁcally want in/out are appropriate for informal use only (Marckwardt and Walcott
1938:49; Shaw 1975:252; Quirk et al 19852732; RH 2“d ed., s.v. ‘want in or out’;
Greenbaum & Whitcut 19882762; OED 2“d ed., s.v. ‘want’; K. Wilson 1993; OADLG
199921140). Some of these sources state that W+PA is very informal and/or not

appropriate for formal use; for example, in the Longman Guide to English Usage, want

46

 

in/out are ‘rather informal’ (Greenbaum & Whitcut 19882762). The Columbia Guide to
Standard American English goes down a few notches from this by stating that forms of
W+PA are ‘casual at best, but more likely nonstandard’ (K. Wilson 1993). A few sources
reject W+PA for formal use altogether (Leonard 19322124; Bryant 19622224): ‘[want in]
has no standing for more formal or general use’ (Leonard 19322124).

It is not clear what the claims about the use of W+PA (with the exception of
Leonard 1932) are based on. What is clear is that further research is needed to accurately
characterize the stylistic distribution of W+PA. Is W+PA used only or even primarily in
informal settings? Finegan and Biber (1994) propose an approach to style or register that
is based on the features of linguistic economy and elaboration. Linguistic forms that show
economy or are compressed, i.e. shorter, are linked to informal styles, while those forms
that exhibit elaboration or are more explicit, i.e. longer, are tied to formal styles (Finegan
& Biber 1994:320-321). Under this view, one would expect W+PA‘to be more common
in informal styles, since W+PA could be considered a more economical form than W +
to-inﬁnitive + PA, which could be considered more elaborate. In the face of little other
evidence, I hypothesize that W+PA and related forms are more commonly and perhaps
exclusively used in informal styles.

Furthermore, sociolinguists, particularly variationists, have typically operated
under the assumption that there is a direct relationship between the distribution of a
linguistic variable across socioeconomic classes and the distribution of that same variable
across styles (e.g. Romaine 19802228; Labov 19642101-102, 1966:222-223, 19722126;
Bell 19842146; Finegan & Biber 19942315-316; Wolfram & Fasold 1974:85-86). Three

basic patterns of the intersection of social and stylistic distribution of a linguistic variable

47

are well-documented in the literature (see, e.g. Labov 1972:178-179, 23 8-244; Bell
1984:151-153; Downes 1998:187-190; Wolfram & Schilling-Estes 19982161-162, 218-

221)

(1) Sociolinguistic markers: Linguistic variables stratiﬁed by both social and stylistic
variables, as is the case with (ing) in English, as shown in Figure 2.5 (Labov 1972223 8-

242; Downes 19982187-190; Wolfram & Schilling-Estes 19982162, 219).

90% -~,

80% i .

 

l- - - Lower

— - - Working
— Lower middle
I—.- Upper middle

 

 

 

 

0% r » ——-

Reading style Careful speech Casual speech

Figure 2.5. Socioeconomic and style stratiﬁcation of (ing) for European-American adults
in New York City (from Labov 19662398, 19722239).

(2) Sociolinguistic indicators: Linguistic variables stratiﬁed by social variables but

not by style, as is currently the case with the Northern Cities Chain Shift in the United

48

States (Labov 19722178; Downes 19982187-190; Wolfram & Schilling-Estes 19982162).

A hypothetical example of a sociolinguistic indicator is given in Figure 2.6.

 

 

 

 

 

 

40% I
I
I. ..................................
|
30% i
I
i— ——————————————————————— . ‘1
. - - - Lower working:
— - - Upper working
20? . .
0 Lower middle l
——Upper middle f
10%
{— __________________
J
0% i- ——— A ————— ——- ———— #—~ .
Formal lnforrnal

Figure 2.6. Hypothetical case of socioeconomic stratiﬁcation with no stylistic variation

(3) Hypercorrection: Linguistic variables stratiﬁed by both socioeconomic class and
style, of which one socioeconomic class’s use (typically the upper working or lower
middle class) will crossover that of another class (typically the upper middle class), as
illustrated in Figure 2.7 (e.g. Labov 1972:124-127; Downes 1998:188-196; Wolfram &

Schilling-Estes 19982219-220).

49

70% 7

 

  
   

 

60%
50% -
, — —Upper middle
40% I Lower middle
--- Upper working
30% -- - -Lower working
- - - Lower

 

 

 

20% 1’

10% i

0% L -- . . “MM".

Word lists Reading style Careful speech Casual speech

 

Figure 2. 7. Socioeconomic and stylistic stratiﬁcation with hypercorrection of (r) 1n adults
in New York City (adapted from Labov 19722114)

Although rarely mentioned in sociolinguistic literature, a fourth pattern could be
possible: Linguistic variables stratiﬁed by style but not by socioeconomic class. It has
become widely accepted that a pattern of stylistic without socioeconomic variation is not
only rare but impossible. Bell’s style axiom, in fact, rules out the possibility of stylistic
variation without social variation with its assumption that social variation underlies
stylistic variation: ‘Variation on the style dimension within the speech of a single speaker
derives from and echoes the variation which exists between speakers on the “social”
dimension’ (Bell 19842151). Bell is not alone in believing that stylistic variation is not

possible without social variation; in fact, the widely accepted types of sociolinguistic

50

variables—stereotypes, markers, and indicators—do not include a category for variables
with stylistic but little or no socioeconomic variation (see e.g. Labov 1972:178-179;
Wolfram & Fasold 1974283-87; Downes 19982188-196; Wolfram & Schilling-Estes
19982161-162). Nevertheless, Bell gives an example of such a case, the raising of /ael/ to
nonstandard [el] in Tehran Persian (Modaressi-Tehrani 1978), but refers to it as an
‘extreme’ ‘hyperstyle variable’ (Bell 19842155). It seems then that it is possible for a
linguistic feature to show little or no socioeconomic variation, as I have hypothesized for
W+PA and related constructions in Section 2.3.2, and yet show stylistic variation, but for
Bell (and others) it would have to be an ‘extreme’ variable, and that would not seem to be

the case for the variables under consideration here.

2.7. Hypotheses

To close the chapter, I list the hypothesis, motivated by the previous literature or gaps

therein, that are tested in this study.

(1) The geographic distribution of NIW+PP and N/W+PA will be primarily Midland.

(2) Abstract senses of N/W+PP and N/W+PA will be more widespread, while
concrete senses will be more regional, speciﬁcally Midland.

(3) Within each dialect region, the use of abstract senses of N/W+PP and N/W+PA
will be greater than that of the concrete senses, particularly in the North and the
South.

(4) The use of want will be greater than the use of need in PP and PA constructions.

(5) PA constructions will be more acceptable than PP constructions.

51

(6)

(7)

(8)

(9)
(10)

There will be no rural-urban differences in the distribution of NIW+PP and
hUVV+IbA.

There will be no differences in the use of N/W+PP and N/W+PA between men
and women.

In the Midland, the use of N/W+PP and N/W+PA will not be stratiﬁed by
socioeconomic class.

In the Midland, the use of N/W+PP and N/W+PA will not be stratiﬁed by age.

In the Midland, NIW+PP and N/W+PA will be used more in more informal styles

than in more formal styles.

52

3. Methodology

In this chapter, the methodology of the present study is laid out. Section 3.1 discusses the
linguistic variables of N/W+PP and N/W+PA included in this study. Section 3.2 justiﬁes
the method of data collection (Section 3.2.1) and describes the survey instruments
(Section 3.2.2). Section 3.3 outlines the pool of respondents and the non-linguistic
variables. Finally, Section 3.4 explains the handling of the raw data and the statistical

analyses.

3.1. Linguistic variables

The primary focus of this study is on forms N/W+PP and N/W+PA. Many scholars
assume that these constructions are identical to constructions with a to-inﬁnitive between
the verb and the prepositional phrase/adverb. Moreover, most descriptions of N/W+PA
reﬂect a belief in the primacy of the forms with the inﬁnitive verbs, e. g. want + to-
inﬁnitive + prepositional adverb. Some refer to these forms as ‘elliptical’ (e.g. Griebling
194725; Marckwardt 1948; Cassidy & Hall 1996:769). Others describe NIW+PA in terms
of want + to-inﬁnitive + prepositional adverb (Prettyman 1907276; Pollard 19152243;
Krapp 19252256; Krumpelmann 1939:156; McAtee 1942270; Evans & Evans 19572545;
Dakin 1966b2513; OED 2“d ed., s.v. ‘want’), e.g. ‘want, used without an inﬁnitive
following’ (Pollard 19152243). One of the dangers of this approach is the assumption of
identical meaning, i.e. referential or representational sameness, which is said to be a
necessary condition for the study of variants of a linguistic variable (Lavendera 1977,
1988; Labov 1978). It is impossible and, in this case, irrelevant to establish the semantic

sameness of N/W+PP and need/want + to-inﬁnitive + prepositional phrase as well as

53

N/W+PA and need/want + to-inﬁnitive + prepositional adverb. It is impossible because
native speaker intuitions indicate that NIW+PP and NIW+PA are necessarily more
general than need/want + to-inﬁnitive + prepositional phrase and need/want + to-
inﬁnitive + prepositional adverb. The examples in 1-4 will help illustrate the point.
Sentence 1 has at least three possible inﬁnitive forms with their own semantic nuances,

given in 2-4.

(1) I want in the house.
(2) I want to go in the house.
(3) I want to get in the house.

(4) I want to be in the house.

Sentence 2 indicates the speaker’s desire to physically move from the exterior to the
interior of the house; the staying in the interior of the house could just be for a moment
(e. g. to get a book) or for a longer period of time. In sentence 3, the emphasis is on
obtaining entry into the house. In sentence 4, the speaker wishes to be and remain inside
the house, so the emphasis is not on obtaining entry or moving from outside to inside, but
on physically being in the house. In addition to native speaker intuitions, at least one
dictionary hints at the speciﬁcity of W+PA: ‘to desire to come, go, or be: the cat wants
in; wants out of the deal’ (Merriam- Webster, on-line ed., s.v. ‘want’).

The semantic sameness argument, central to the debate between formal linguists
and sociolinguists (e.g. Lavandera 1977, 1988; Labov 1978) is not relevant to this study,

because only the forms N/W+PP and N/W+PA are examined. In this study, N/W+PA and

54

need/want + to-inﬁnitive + prepositional adverb, for example, are not treated as variants
of a single variable. The issue of whether these constructions are indeed variants of
need/want + to-inﬁnitive + prepositional phrase and need/want + to-inﬁnitive +
prepositional adverb must be left to future study.

As discussed in the previous chapter, a wide range of prepositions and
prepositional adverbs may occur in the grammatical constructions N/W+PA and
N/W+PP. A subset of the possible forms have been selected for this study, since any
potential survey instrument would be prohibitively long if all possible realizations of
these constructions were investigated. Items 5-8 are the forms of N/W+PP that are

included in this study:

(5) need in + NP
(6) need off + NP
(7) want in + NP

(8) want off + NP

Items 9-18 are the forms of N/W+PA that are included in this study:

(9) need by
(10) need down
(1 1) need in
(12) need off

(1 3) need out

55

(14) want by

(1 5) want down
(16) want in
(17) want off

(18) want out

The pool of experimental sentences used in the survey instruments included two tokens
of each of 5-18, all in concrete senses, and two tokens of abstract senses for 5-8, 11-13,
and 16-18. Only concrete senses of need by/down (9-10) and want by/down (14-15) were

used because these forms do not lend themselves well to abstract senses.

3.2. Survey instruments

3.2.1. Study of grammatical features

The study of grammatical features has played the ugly stepsister role to the study of
phonological features in the ﬁeld of sociolinguistic variation. While it is relatively easy to
gather data for the study of phonological variation, and there are many tried and true
methods for eliciting pronunciation at different stylistic levels (e.g. informal {-9
formal), the same cannot be said for the study of grammatical variation. Most studies of
grammatical variation admit to the difficulty of collecting quantitative data, and owing to
this difﬁculty, few move beyond description of the form to attempt any kind of
sociolinguistic stratiﬁcation or stylistic differentiation. Questionnaires, however, are a

good way to collect data on grammatical features.

56

Self-reports (also known as introspective survey instruments) have been
condemned by some sociolinguists as unreliable (e.g., Labov 19752106-108; McDavid
1983) even though they have been commonly used, particularly in dialectology (e. g.
Frazer 1993; Murray 1993; Frazer et a1 1996; Bailey et al 1997; Albanyan & Preston
1998; Murray & Simon 1999, 2002). Interestingly, the evidence against the use of self-
reports is largely anecdotal, which is in stark contrast to the type of data sociolinguists
normally employ. Although observed linguistic behavior in a number of situations or
registers may be the preferred means for gathering phonetic/phonological data, the
limitations of such data collection for relatively rare grammatical constructions (i.e.,
relative to a phonetic/phonological phenomenon) are well recognized. There is, however,
quantitative evidence suggesting that data from self-reports are more valid and reliable
than previously thought. In a study comparing the results of ﬁve linguistic surveys with
differing methodologies, Bailey et al (1997258) found that for some non-
phonetic/phonological features, self-reports are ‘better measures of linguistic behavior
than observations of usage’. There is no doubt that additional research must be done in
this area. In light of Bailey et a1 (1997) and the precedent established in the ﬁeld of
dialectology, I am conﬁdent that for the linguistic variables in this study, self-reports

provide a valid measure of actual use.

3.2.2. Questionnaires
A self-report questionnaire was designed to determine not only the use of the forms of
NIW+PP and N/W+PA by respondents but also their preferred stylistic level. A total of

62 linguistic items were on each version of the questionnaire: 24 experimental items

57

involving a form of N/W+PP or N/W+PA and 38 ﬁllers. The experimental items on each
questionnaire consist of one token of each form of N/W+PP and N/W+PA, including
abstract senses, listed in Section 3.1. The ﬁllers consist of 23 sentences that include other
regional and social dialect forms (e.g. I bought myself a new TV), with an emphasis on
forms that are common in the Midland, and 15 sentences that are acceptable in all
varieties of American English (e. g. Eddie is very athletic). The items that are presumed to
be acceptable in all varieties of American English were included as a way to check
respondents’ performances on the questionnaire. Appendix A contains an organized list
of questionnaire items.

Respondents rated their use of these items on the following forced-choice scale,

taken from Albanyan and Preston (1998):

Choose a if you would never use this sentence.

Choose b if you would use this sentence only in writing or in very formal speech
situations like a job interview or in court.

Choose c if you would use this sentence only with close friends and/or family.
Choose (1 if you would use this sentence in general conversation, in stores and with
people you don’t know well.

Choose e if you would use this sentence in all situations.
This scale was chosen because it allowed for the investigation of two distinct

judgments: 1) grammaticality or use and 2) stylistic differentiation. A response of ‘a’

indicated that the respondent would not use the sentence, while a response of ‘b’, ‘c’, ‘d’,

58

or ‘e’ indicated that the respondent would use the sentence. Responses ‘b’, ‘c’, ‘d’, and
‘e’ also revealed the style or register in which the sentence was judged to be most
appropriate. Responses ‘a’ through ‘e’ do not represent _a continuum and are not treated
as such.

Four versions of the questionnaire, labeled A1, A2, B1, B2 were created.
Versions A1 and A2 had exactly the same items on them; however, the order of items had
been scrambled to counteract any ordering effects. The same is true of versions B1 and
B2. Although the A and B versions of the questionnaires used the same ﬁllers, they
contained different experimental items, listed in Appendix A. One version of the
questionnaire (A1) is given in Appendix B. The questionnaire was pilot tested on a small

number of people in Michigan and Ohio before it was used to gather data for this study.

3.3. Respondents and non-linguistic variables

A total of 163 respondents’ questionnaires were used for data analysis'. There were two
criteria for inclusion in this study: (1) Respondents had to have grown up in the United
States east of the Mississippi River; and (2) Respondents had to have grown up in the
same area between the ages of 2-15. The majority of respondents—148—were university
and technical college students enrolled in institutions in Ohio, Michigan, and Georgiaz.
Limited demographic data, e.g. age, ethnicity/race, sex, hometown, and occupation of

primary wage-eamer in the family, were obtained from these respondents. The other 15

 

' 26 questionnaires (in addition to the 163 used for data analysis) had to be thrown out because they were
incomplete (10) or the respondents did not match the criteria for inclusion in the study: 10 were raised
outside of the US; 3 were raised in the western United States; and 3 moved to different states several times
as children.

2 l gratefully acknowledge Bridget Anderson, Gord Easson, Peggy Ketner, Nevin Leder, and Patrick
Mullen for distributing questionnaires to their students.

59

respondents were residents of a cluster of small cities in central Ohio, east of Columbus,
and were recruited for this study through local churches. These 15 respondents completed
the questionnaires and gave additional demographic data, including their family
background and their orientation to the local area. A list of all respondents is given in
Appendix D.

Based on the demographic data solicited on the questionnaire, respondents were

classiﬁed into several regional and social groups, as shown in Table 3.1.

Table 3.1. Demographic distribution of respondents (see Appendix D)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Total North Midland South

Total 163 41 95 27
Rural-Urban

Rural 42 (26%) 8 (20%) 28 (29%) 6 (22%)

Urban 121 (74%) 33 (80%) 67 (71%) 21 (78%)
Gender

Male 114 (70%) 31 (76%) 64 (67%) 19 (70%)

Female 49 (30%) 10 (24%) 31 (33%) 8 (30%)
Socioeconomic Status

Upper Middle 50 (31%) 17 (41%) 21 (22%) 11 (41%)

Middle Middle 28 (17%) 6 (15%) 14 (15%) 8 (30%)

Lower Middle 51 (25%) 11 (27%) 24 (25%) 6 (22%)

Upper Working 25 (15%) 2 (5%) 21 (22%) 2 (7%)

Lower Working 12 (7%) 1 (2%) 11 (12%) 0 (0%)

No response 7 (4%) 3 (7%) 4 (4%) 0 (0%)
_A_ge

18-24 years 128 (79%) 40 (98%) 64 (67%) 24 (89%)

25-34 years 15 (9%) 1 (2%) 11 (12%) 3 (11%)

35-54 years 10 (6%) 0 (0%) 10 (11%) 0 (0%)

55+ years 10 (6%) 0 (0%) 10 (11%) 0 (0%)
Race/Ethnicity

European American 137 (84%) 33 (80%) 84 (88%) 20 (74%)

Other 19 (12%) 5 (12%) 9 (9%) 5 (19%)

No response 7 (4%) 3 (7%) 2 (2%) 2 (7%)

 

60

 

On the survey instrument, presented as Appendix B, respondents were asked to write
down the city and state where they grew up between the ages of two and ﬁfteen (Item
67). This information was used to classify them as belonging to one of three dialect
regions: the North, the Midland, or the South, deﬁned in Figure 2.1. Since the focus of
this study is on Midlanders’ use of N/W+PP and N/W+PA, Midland residents were
sought out; thus, the greatest number of respondents, 58%, were from the Midland. The
respondents from the North and South were used, in effect, to provide a comparison for
Midland use, so fewer respondents were from these areas: 25% from the North and 17%
from the South, as shown in Table 3.1.

Respondents were also grouped according to four non-linguistic variables: rural-
urban residence, socioeconomic status, age, and gender. Respondents were classiﬁed into
the categories of rural and urban residence based on their hometown, given in response to
questionnaire item 67. Three criteria were used to deﬁne rural and urban areas:
population, distance from major metropolitan area, and location on or near a major
highway. As deﬁned in this study, rural areas have a population of less than 10,000, are
more than 10 miles from a major metropolitan area, and are not located on a major
highway. All other areas were included in the urban category. Table 3.1 displays the
breakdown of respondents by rural-urban residence.

The second social variable in this study is gender. Respondents were grouped into
two categories based on a self-report of gender, in response to questionnaire item 64. In
Table 3.1, the distribution of respondents across gender categories is given.

Based on the occupation of the primary wage earner in the family, given in

response to questionnaire item 68, respondents were divided among ﬁve socioeconomic

61

classes: upper middle, middle middle, lower middle, upper working, and lower working.
This study is not the ﬁrst to base social class assignment on occupation alone (see c. g.
Pederson 19912xvii-xviii; Chambers 1995246-47). The scale used for the purposes of
social class assignment is largely based on Warner’s revised scale (1960:140-141). The
passage of time and the creation of new technology/jobs necessitates updating Warner’s

scale:

The most important fact to remember about using the ISC as a measurement of
social class is that, in order for it to be a reliable instrument and accurate index of
social class, . . . [the] scales must reﬂect how Americans feel and think about the
relative worth of each job . . . (Warner 1960240).

Although sociological surveys have generally found ratings of occupational prestige to be
quite stable over time, they have also demonstrated that occupational prestige does
change (Hodge et a1 1964; Nakao & Treas 1994). Therefore, I have updated Warner’s
scale using data from a study of public opinion on occupational prestige (Nakao & Treas
1994). The occupations listed in Table 3.2 are representative of those belonging to each

category.

62

Table 3.2. Occupational prestige ratings

 

Rating

Representative Occupations

 

1

Judges, Lawyers, Architects, Aerospace and Chemical Engineers, Chemists,
University Professors, Dentists, Physicians

 

2

Accountants, Primary and Secondary School Teachers, Police Ofﬁcers,
Professional Clergy, Computer Programmers, Industrial and Mechanical
Engineers, Agricultural Scientists, Optometrists, Psychologists, Registered
Nurses, Pharmacists, Veterinarians

 

Electrical Technicians, Journalists, Librarians, Nursery School Teachers, School
Counselors, Dental Hygienists, X-Ray Technicians, Fireﬁghters, Farm Owners

 

Real Estate Agents, Insurance Agents, Public Relations Specialists, Lay Clergy,
Mail Carriers, Bank Tellers, Secretaries, Grain/Cattle Farmers, Electricians, and
Construction/Factory Forepersons

 

Security Guards, Prison Guards, Survey and Mapping Technicians,
Receptionists, Shipping Clerks, Restaurant Cooks, Hair Stylists, Cosmetologists,
Child Care Workers, Automobile Mechanics, Telephone Installers/Repairers,
and Skilled Construction Craftsmen

 

Used Car Salespersons, Retail Sales Associates, Cashiers, Bill Collectors,
Janitors, Housekeepers, Fast Food Employees, Restaurant Servers, Factory
Workers, Truck Drivers, and Garbage Collectors

 

 

 

Newspaper deliverers, Migrant Workers, Envelope Stuffers, Dishwashers, and
Grocery Baggers

 

The socioeconomic classes in this study correspond to Warner’s scale in the

following way: Warner’s 1-2 is upper middle class; Warner’s 3 is middle middle class;

Warner’s 4 is lower middle class; Warner’s 5 is upper working class; Warner’s 6-7 is

lower working class. In Table 3.1, the breakdown of respondents by social class is given.

For the purposes of studying age as a social variable, respondents were classiﬁed

into four age groups, based on their response to item 65: 18-24 years old, 25-34 years old;

35-54 years, and 55 years old and over. These age classiﬁcations are not simply arbitrary

classiﬁcations based on the etic approach, which is typical of quantitative sociolinguistic

studies, but are an attempt to appeal to the emic approach which ‘groups speakers

according to some shared experience of time’ including external events (Eckert 1997).

63

 

These age classiﬁcations are based on the similar stages of life that individuals in these
age ranges ﬁnd themselves.

Those in the 18-24 year old range are beginning adult life, living on their own for
the ﬁrst time, completing their formal education, etc. Those in the 25-34 year old range
are typically beginning families, starting their careers, and buying their ﬁrst homes.
Those in the 35-54 year old range are typically established in their careers and in their
families, though some may be experiencing career and family changes. Those in the 55+
years group are nearing retirement or already retired, are most likely living alone or with
a signiﬁcant other; and if they have children, the children are grown and living on their
own. The majority of respondents in the entire sample fell into the 18-24 years group,
79% (128). The other three age groups have fewer members but are roughly equivalent

in size, as shown in Table 3.1.

3.4. Data handling and statistical analyses
In most of Chapter 4, in particular Sections 4.1-4.4, the data are viewed categorically in
terms of ‘non-use’ and ‘use’. Only ‘use’ is reported on, and it is deﬁned broadly to
include use in formal situations, use with family and friends, use in general situations, or
use in all situations. Thus, ‘b’, ‘c’, ‘d’, or ‘e’ responses (outlined in Section 3.2) to a
questionnaire item equal ‘use’. In the statistical tests reported in Sections 4.1-4.4,
responses ‘b’, ‘c’, ‘d’, and ‘e’ were collapsed. In Section 4.5, these responses are treated
individually.

The statistical results presented in Chapter 4 are all based on x2 tests of

signiﬁcance. A number of preliminary tests were run, however, to prepare the data for the

64

statistical analyses presented in the next chapter. Because two different token sentences
were used for each form of N/W+PP and N/W+PA on questionnaires A and B, it was
necessary to run tests of signiﬁcance to determine if these token sentences were treated
the same by respondents. x2 tests determined that there were no signiﬁcant differences
between the tokens of N/W+PP and N/W+PA on questionnaires A and B, and the results
of the two questionnaires were conﬂated, with a few exceptions, discussed in Section 4.1.

Originally, respondents were classiﬁed into ﬁve geographic categories based on
their hometown: North, North Midland, Mid Midland, South Midland, and South. The
three Midland categories were conﬂated when no signiﬁcant differences were found to
exist among them.

In the x2 tests of signiﬁcance presented in Chapter 4, missing responses were
omitted in the running of the statistical tests, and the values of statistical signiﬁcance
given are actual values rounded to three decimal places. In the data tables, please note the
following: 1) ‘N’ equals the total number of tokens of a given item (not the number of
respondents); 2) the use of a linguistic variable is given as a rounded percentage followed
by the number of responses in parentheses. Having covered the background of N/W+PP
and N/W+PA and the methodology, the focus now shifts to the results and discussion of

the raw data and statistical analyses in Chapter 4.

65

4. Results and discussion
In this chapter, results are presented and discussed within each section. In section 4.1,
usage statistics of N/W+PP and N/W+PA in three dialect areas, the North, the Midland,
and the South, are given. Section 4.2 examines the relevance of the semantic distinction
between concrete and abstract in the distribution of N/W+PP and N/W+PA across and
within the dialect regions. In section 4.3 usage hierarchies among the various forms of
N/W+PP and N/W+PA within each dialect region are presented, i.e. need vs. want and
PP vs. PA. Section 4.4 looks at nonlinguistic factors, i.e. rural-urban residence,
socioeconomic status, age, and gender, relating to the use of N/W+PP and N/W+PA
across and within the dialect regions. Finally, Section 4.5 examines the stylistic
distribution N/W+PP and N/W+PA.

The reader is reminded that in the tables and ﬁgures, all usage results are rounded
to whole numbers; the results of tests of statistical signiﬁcance are rounded to three

deCimal points; and ‘N’ is the number of tokens of a given item.

4.1. Use across dialect regions

In this section, the usage statistics for N/W+PP and N/W+PA are presented for the North,
the Midland, and the South (recall that ﬁner distinctions within the Midland were not
signiﬁcant). The focus of this section is on the use of the constructions across the dialect
regions. These results should in no way be construed as representing a complete mapping
of these constructions in the North, the Midland, and the South, let alone the entire
United States; rather, they are intended to provide a better understanding of the use of

these constructions in the Midland, and in order to do that, some data from the North and

66

South is needed. A comparison of the use of these constructions within each region, for
example, the difference between the use of the forms of N+PP versus W+PP, is pursued

in detail in Section 4.3.

4.1.1. NIW+PP
As discussed in Section 2.2.1, N+PP, as in 1-4, and W+PP, as in 5-8, have previously

received very little attention.

(1) Dean needs in the building. (1)

(2) Katy needs in this class in order to graduate this semester. (4)
(3) Peter needs off the plane. (5)

(4) The lawyer had a conﬂict of interest and needed off the case. (8)
(5) When the cat wants in the house, she scratches at the door. (10)
(6) This student wants in your class. (1 1)

(7) Do you want off the merry-go-round? (14)

(8) I want off your mailing list. (16)

Nevertheless, the use of N/W+PP is robust in some parts of the United States. The
percentage use of N/W+PP by dialect region is presented in Figure 4.1. In the Midland,
both N+PP and W+PP are in common use. Although the use of N+PP is less in the North
and South, it is not so low as to be nonexistent. In fact, the use of W+PP in the North and
South is well over 50%. The distribution of N+PP and W+PP, however, is signiﬁcant

across the dialect regions.

67

100% 1.
90% . 390
80% «
70% — 70%
60% —
50% l

40% ‘ 41%

 

 

30% i
20% l
10% 1

 

0% +—— —— —- . . ———— ~-
North (N=l64) Midland (N=380) South (N=108)
I ' +W+PP +N+PI37

Figure 4.1. Use of N+PP (Pearson x2=98, df 2, p=.000) and W+PP (Pearson x2=56, df 2,
p=.000) by dialect region

 

 

The use of the forms of N+PP and W+PP are presented in Table 4.1. The use of
all forms—need in + NP, need oﬂ+ NP, want in + NP, want oﬂ+ NP—is signiﬁcantly
different across the dialect regions. In each case, the use is greatest in the Midland. It
should be mentioned that in the Midland, the concrete token items for need in + NP on
the two questionnaires (items 3 and 4 in Appendix A) were rated signiﬁcantly differently
from each other (Yates x2=8, (if 1, p=.005); however, in spite of the signiﬁcance, neither
rating could be considered an outlier, and throwing out either item did not affect the
overall results. The use of the forms of N+PP and W+PP are generally slightly greater in
the North than in the South. The hypothesis of a mainly Midland distribution of N+PP

and W+PP is supported.

68

Table 4.1. Use of N/W+PP by dialect region

 

N . Use 2 Pearson

 

 

 

 

xidrlh

 

need in + NP

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

North 82 37% (30)
Midland 189 ‘_ 77% (145) j: 54 2 .000
South 54 35%(19)

need off +NP
North 82 45% (37)

Midland 190 g, 78% (149) 44 2 .000
South 54 39% (21)

want in+NP
North 82 57% (47)

Midland 190 385%(162)  39 2 .000
South 54 50% (27)

want off + NP
North 82 83% (68)

Midland 19o '93%(l77)7 21 2 .000
South 54 70% (38)

 

 

 

 

 

 

Also noteworthy are the differences in the use of the forms of N+PP versus

W+PP, particularly in the North and South. As shown in Table 4.1 need in + NP and

need oﬂ+ NP are the two least used forms in all three dialect areas, while want in + NP

and want oﬂ+ NP are the most used. These differences point to usage hierarchies

between need and want, which are examined in greater detail in Section 4.3.

Nevertheless, it is also notable that want ojf+ NP is the most used form in all three

dialect areas, and the use of this form in the North and South is substantially greater than

any of the other three forms.

69

 

4.1.2. N/W-l-PA

Similar to N/W+PP, N+PA, as in 9-13, has received little or no attention; however,

W+PA, as in 14-18, has received the most attention of any of the constructions of interest

here.

(9)

(10)
(11)
(12)
(13)
(14)
(15)
(16)
(17)
(18)

I’m sorry to bother you, but I need by. (17)

The cat is stuck in the tree and needs down. (20)

The day care center is full and my child needs in. (23)

The fumes from this bus are making me sick. I need off. (25)

The company agreed to a deal two weeks ago, but now they need out. (32)

I can move if you want by. (34)

The baby wants down. (36)

The kids say it’s too hot outside, and they want in. (3 7)

The baseball team hasn’t won a game in several weeks, and Jim wants oﬁ‘. (43)

This marriage is over. I want out. (48)

Figure 4.2 displays the overall use of N/W+PA across the dialect regions. The regional

distribution of N+PA and W+PA is signiﬁcant. Use of N+PA and W+PA in the Midland

(81%, 84%) outpaces that in the North (47%, 68%) and the South (50%, 69%). Use in the

North and the South is considerably less than that in the Midland; however, use in the

North and the South is similar.

70

100% '1

90% i 84%
80% 1
70% A 68% 81% ' 69%
60% l
50% ~

0
40% i 47% 50/0

 

30% l
20% l

10% 1

 

ml . ..
North (N=328) Midland (N=713) South (N=216)
I —D— W+PA —A—N+PA_j

Figure 4.2. Use of N+PA (Pearson x2=147, df 2, p=.000) and W+PA (Pearson x2=41, df
2, p=.000) by dialect region

 

 

The regional distribution of the forms of N+PA is shown in Table 4.2. For all
forms of N+PA, the use across the dialect areas is signiﬁcant (p=.000). The use of each
form is greatest in the Midland. It should be noted that in the Midland, the concrete token
items for need of on the two questionnaires (items 25 and 26 in Appendix A) were rated
signiﬁcantly differently ﬁ'om each other (Yates x2=6, df 1, p=.015); the ratings for item
26 were outliers and were thrown out because they were most inconsistent with other
results. Generally, the use in the South is slightly greater than in the North, except in the
case of need in, which reverses this pattern. The hypothesis that the use of the forms of

N+PA is greatest in the Midland is well supported by the data.

71

Table 4.2. Use of N+PA by dialect region

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pearson
N Use x; I df l P
need by .
North 41 34% (14)
Midland 95 80%(76) ‘ 29 2 .000
South 27 48% (13)
need down
North 41 39% (16)
Midland 95 82% (78) .. 29 2 .000
South 27 44% (12)
need in
North 82 55% (45)
Midland 190 86%1163) f 45 2 .000
South 54 48% (26)
need off
North 82 48% (39)
Midland‘ 144 l 176%(110)j 21 2 .000
South 53 55% (29)
need out
North 82 48% (39)
Midland 188 _‘ 78%,(147). 4' 30 2 .000
South 54 52% (28)

 

 

 

 

 

 

The behavior of the forms of N+PA allows two observations. One is that the
range of use of the various forms of N+PA is quite narrowly clustered, particularly in the
Midland and the South. In the Midland, the distance between the least used form, need 017
(76%), and the most used form, need in (86%), is 10%, and in the South, the distance
between the least used form, need down (44%), and the most used form, need oﬂ(55%) is

11%. This contrasts sharply with the distance between the least used and most used forms

 

' In the Midland, the concrete token items for need 0,?" on the two questionnaires (items 25 and 26 in
Appendix A) were rated signiﬁcantly differently from each other (Yates xz=6, df l, p=.015); item 26
ratings were outliers and were thrown out.

72

 

of NIW+PP. Another observation is that the forms of N+PA, unlike the forms of
N/W+PP, behave differently in the Midland than in the North and South. For example,
the least used forms in the Midland are need of and need out; however, in the North and
South the two least used forms are need by and need down.

The usage statistics for the forms of W+PA are given in Table 4.3. As with the

forms of NIW+PP and N+PA, the Midland continues to dominate in the use of the forms

Table 4.3. Use of W+PA by dialect region

Pearson
df

want
North
Midland
South
want down
North
Midland
South
want in
North 74%
Midland 85%
South 76%
want off
North 65%

South 69%
want out

North 66% 54

Midland

South 63% 34

 

 

2 In the Midland, the abstract token items for want of on the two questionnaires (items 43 and 44 in
Appendix A) were rated signiﬁcantly differently ﬁ'om each other (Yates x2=14, df 1, p=.000); item 44
ratings were outliers and were thrown out.

73

of W+PA acrosspthe dialect regions; however, a different picture emerges than seen thus
far. The distribution across the dialect regions is signiﬁcant for want by (p=.003 ), want
off (p=.001), and want out (p=.011), but it is not signiﬁcant for want down (p=.111). It
bears mentioning that in the Midland, the abstract token items for want 017 on
questionnaires A and B (items 43 and 44 in Appendix A) were rated Signiﬁcantly
differently from each other (Yates xz=14, df 1, p=.000); the ratings for questionnaire item
B were outliers and were thrown out because they were most inconsistent with other
results.

While the distribution of want in approaches signiﬁcance, further clariﬁcation of
this form is warranted. Among Northern respondents, the abstract token sentences for
want in on the two questionnaires (items 39 and 40 in Appendix A) were rated
Signiﬁcantly differently from each other. Item 39 (45%) was rated much less acceptable
than item 40 (95%). This pattern ran contrary to comparisons with other token sentences
of want in as well as other forms of W+PA; however, neither item by itself patterned
consistently with other forms and both were maintained. Only the results from want by,
want oﬁf want out, and possibly want in support the hypothesis of a stronger Midland
distribution for the forms of W+PA.

As seen with the forms of N/W+PP, there is a wide range of use between the least
and most used forms of W+PA in each dialect region. For example, in the North, the least
acceptable form, want by is used by 51% of respondents, while the most used form, want
down is used by 85% of respondents. In addition, the use of the forms of W+PA is quite
consistent across the dialect regions: The order from least to most used in all three

regions is want by, want out, want oﬁf want in, and want down.

74

4.1.3. Discussion

Several intriguing ﬁndings about the use of NIW+PP and N/W+PA in the North, the
Midland, and the South are further discussed in this section. In particular, this section
focuses on four major ﬁndings: 1) The robust Midland distribution of NIW+PP and
N/W+PA, supporting the hypothesis; 2) the greater than expected use of N/W+PP and
N/W+PA in the North and South; 3) the greater than expected use based on the lack of
previous literature of prepositional phrase forms and need forms; 4) the differences in the
use of individual forms within and across the dialect regions.

The concentration of use of NIW+PP and N/W+PA in the Midland, with the
exception of want down and want in, is in line with the hypothesis of a more regional,
speciﬁcally Midland, distribution of these constructions. The strong Midland distribution
of these constructions is similar to the distribution of other grammatical constructions
discussed in Section 2.3 including need/want/like + past participle and positive anymore
(Youmans 1986; Murray 1993; Frazer et a1 1996; Murray & Simon 1999, 2002). The
Midland, however, is not alone, in its use of NIW+PP and NIW+PA.

In the North and the South, the use of N/W+PP and N/W+PA is greater than
previous literature would lead one to expect. The use of some forms of N/W+PP and
NIW+PA exceeds 50%: The use of all forms of W+PP and W+PA is greater than 50% in
the North and the South; in addition, the use of need in in the North and need of and need
out in the South is greater than 50%, albeit Slightly. Such results bring into question the
true status of these constructions in United States. As discussed in Section 2.3, the data
from dialect word lists, dialectology projects and studies, and usage guides, suggested

that these constructions were uncommon to nonexistent in the North and South. Perhaps

75

the best illustration of this is found in a statement from Kurath (194930): ‘The records of
the Linguistic Atlas Show that I want off is in common use in nearly all of the Midland
[. . .] and that it is not current at all in the North and the South’.

What accounts for the levels of use of N/W+PP and N/W+PA found in this study?
One possibility is that generalizations such as Kurath’s were based on too little evidence.
This argument may apply to the forms of NIW+PP and N+PA; however, the North and
South were richly sampled for the use of want of in LANCS, LAUM, LAGS, and DARE
and few attestations were found in the North, in particular, and in the South (see Figures
2.2 and 2.3). A second possibility has to do with a learning curve effect. Perhaps because
there were so many tokens of N/W+PP and NIW+PA on the questionnaire, respondents
were inﬂuenced by this and grew more and more accustomed to the constructions while
completing the questionnaires. Statistical analyses of the results revealed no evidence of a
learning curve; see Appendix E for additional details. A third possibility is that these
forms have developed independently in the Midland, the North, and the South, and the
emergence of these forms has been more recent in the North and the South than in the
Midland (Hudson, p.c.). This is a possible scenario, which requires additional research.

A fourth and perhaps more likely possibility is that these forms have been
gradually spreading from the Midland to the North and the South, which might explain
the lesser degree of use in the North and the South reported in the early Linguistic Atlas
projects and DARE data as compared to the somewhat greater use found in LAGS (see
Figures 2-3 in Section 2.3) and much greater use reported in the present study. In
addition, other grammatical constructions with strong Midland distributions are

reportedly encroaching on neighboring dialect areas. For example, studies indicate that

76

the use of need/want + past participle and positive anymore is spreading beyond the
Midland (Labov 1973:72; Youmans 1986:74; Labov 1991:277; Murray 1993:175, 185;
Frazer et al 1996:268; Murray & Simon 2002:57; Murray 2002:349). Furthermore,
discussions with respondents (and my own experience as a native speaker) reveal that
users of N/W+PP and N/W+PA are unaware that these constructions are not used all over
the United States and that they are considered unacceptable by many. These constructions
appear to ﬂoat under the radar of conscious awareness and seem not to have any stigma .
or even negative evaluation attached to their use in the Midland, which if true in other
areas, may facilitate their spread. Additional research on attitudes towards N/W+PP and
N/W+PA is needed.

It bears mentioning that the data Show greater use of NIW+PP and N+PA,
particularly in the Midland, than previous research would lead one to expect. N/W+PP
and N+PA have been virtually ignored in the literature; nevertheless, these constructions
are uSed by a large percentage of Midlanders and not a minute percentage of Northern
and Southerners. It is surprising that W+PA has received so much attention, while these
other seemingly closely related forms have languished in obscurity. As mentioned
previously, I found no usage guides that had entries for N/W+PP and N+PA (with the
exception of DARE), no citations of N+PP, and only a few of W+PP and N+PA. The
scarcity of previous literature on NIW+PP and N+PA resembles the status of
need/want/Iike + past participle prior to the recent work of Frazer, Murray, and Simon
(Frazer et a1 1996; Murray & Simon 1999, 2002). It is not clear why N/W+PP and N+PA
have received little notice; however, what is clear is that these constructions are regularly

used in the Midland, and to a lesser degree in the North and the South, as well.

77

The geographic patterns of use of N/W+PP and N/W+PA, both synchronically
and diachronically, would seem to add to the continuing debate over the status of the
Midland dialect area (e.g. Bailey 1968; Carver 1987; Davis and Houck 1992, 1996;
Frazer 1993, 1994, 1996; Johnson 1994; McElhinny 1999; Flanigan 2000; Flanigan and
Norris 2000). The ﬁndings presented here seem to lend support to the idea of the
Midland as an separate dialect area, rather than simply a transition zone between the
North and the South, particularly if it is assumed that N/W+PP and NIW+PA have their
origins in the Midland (while additional research is needed to substantiate that every form
has Scotch-Irish/Midland American roots, compelling work has already provided this link
for some forms, e.g. Stanley 1936az3-4; Marckwardt 1948:9; Crozier 1984:326;
Montgomery 1997:196), and have begun spreading from the Midland into other areas.
Continued examination into the status and etiology of the Midland dialect area is
warranted.

Finally, striking patterns in the use of NIW+PP and N/W+PA can be seen when
the forms of these constructions are ranked in order of use from least to greatest within
each dialect region, as in Table 4.4. Two patterns are immediately noticed: I) the polarity
of need and want forms, particularly in the North and the South, with the majority of need
forms at one end of the scale and the majority of want forms at the other end; and 2) the
wide range of use between the least used and most used forms, again, particularly in the
North and the South. The preponderance of need forms at the lower end of the usage
scale and want forms at the higher end of the usage scale is especially evident in the
North and the South. In the Midland, the forms of need and want tend to be more evenly

dispersed through the rank order, though even there more want forms are at the higher

78

end than need forms. In addition, the prepositional phrase forms are interspersed with the

prepositional adverb forms in all three areas.

Table 4.4. Rank order of N/W+PP and N/W+PA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

North %use Midland %use South %use
need by 34% need off 76% need in + NP 35%
need in + NP 37% need in + NP 77% need off + NP 39%
need down 39% want by 77% need down 44%
need off + NP 45% need off + NP 78% need by 48%
need off 48% need out 78% need in 48%
need out 48% want out 79% want in + NP 50%
want by 51% need by 80% need out 52%
need in 55% need down 82% want by 52%
want in + NP 57% want in + NP _8_5_% need off 55%
want'Off '1 i . _' I ' I '-‘ I » want out
want out 1: "
want-ins ' '
wantdo . . "

 

 

 

 

These observations lend support to the notion that NIW+PP and N/W+PA are
ﬁrmly entrenched in the Midland and are perhaps in the process of gaining acceptability
in the North and the South. It appears that in the North and the South, forms of W+PP
and W+PA are gaining greater acceptability than N+PP and N+PA. The differences in the
use of need and want forms and PP and PA forms within each dialect region are
examined in greater detail in Section 4.3.

Further support for a change in progress is found in the wide range of
acceptability between the least used and most used forms of N/W+PP and N/W+PA,
particularly in the North and the South. Interestingly, four of the ﬁve most used

forms—want down, want in, want of and want oﬂ+ NP—are the same in the North, the

79

Midland, and the South. In the Midland, want oﬂ and want in are expected in the top ﬁve;
however, the three most used forms want down, want oﬂ+ NP, and need in are
surprising. It is also surprising that want out is not among the most used forms in the
Midland, Since so much of the previous literature focused on this construction. In the
North and the South, the high use of these forms—want down, want in, want off and
want oﬂ+ NP—along with want out suggests that they are perhaps the oldest in the
spread. Finally, this greater variation of use of the forms of NIW+PP and NIW+PA in the
North and the South suggests that if they are spreading, they are perhaps spreading as
lexical items rather than grammatical features. Additional research on the use of the
forms of N/W+PP and N/W+PA in individual communities in the North and the South
with an eye toward change is needed.

In summary, the most Signiﬁcant ﬁndings in this section include the following: 1)
The greater Midland distribution of N/W+PP and N/W+PA; 2) the not insigniﬁcant use
of these constructions in the North and the South; 3) the high rate of use of need and PP
forms of the constructions; and 4) the wide range of acceptability of various forms of
NIW+PP and NIW+PA as well as the similarity of their rankings, especially in the North
and the South. Additional research is needed to better understand these constructions,
especially in the following two areas: 1) to determine whether a change in progress is at
work for N/W+PP and NIW+PA, particularly in the North and the South, and 2) to gain
an understanding as to whether these constructions are negatively, positively, or neutrally

evaluated by speakers in all three dialect areas.

80

4.2. Semantic features

The usage statistics presented in Section 4.1 Show only part of the picture of the
distribution of N/W+PP and N/W+PA. As discussed in Section 2.4, the semantic
distinction of concrete and abstract may play a role in the regional distribution of these
constructions. Recall that sentences like 19-21 illustrate concrete senses, while sentences

like 22-24 illustrate abstract senses.

( I 9) Dean needs in the building. (1)

(20) The kids say it’s too hot outside, and they want in. (3 7)

(21) When the roller coaster picked up speed, Ed yelled that he wanted off. (41)
(22) Katy needs in this class in order to graduate this semester. (4)

(23) That sounds like a great plan. I want in. (40)

(24) The baseball team hasn’t won a game in several weeks, and Jim wants off. (43)

In some cases, concrete and abstract senses of N/W+PP and N/W+PA pattern differently

across and within the dialect regions.

4.2.1. N/W+PP

Table 4.5 presents the usage statistics for concrete and abstract senses of N/W+PP across
the dialect regions. The regional distribution of concrete and abstract senses of the forms
of N/W+PP resembles that of all forms of N/W+PP taken together, shown in Table 4.2;
differences in the use of each form across the regions are signiﬁcant. For each form, use

in the Midland is Signiﬁcantly greater than use in the North and the South.

81

Table 4.5. Concrete and abstract senses of N/W+PP across dialect regions

 

 

N

 

Use

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:5 df 1 J
need in + NP (Concrete) .
North 41 32% (13)
Midland 94 {69%(65‘11 f 27 .000
South 27 22% (6)
need in + NP (Abstract)
North 41 41% (17)
Midland 95 784%}(80jzf 29 .000
South 27 48% (13)
need off + NP (Concrete)
North 41 39% (16)
Midland 95 {72%.(68) 24 .000
South 27 26% (7)
need off + NP (Abstract)
North 41 51% (21)
Midland 95 85%(811) 22 .000
South 27 52% (14)
want in + NP (Concrete)
North 41 68% (28) ,
Midland 95 3.91%;(86)55f 13 .001
South 27 67% (18)
want in + NP (Abstract)
North 41 46% (19)
Midland 95 '1;80%-(i7;o)j;zgz 27 .000
South 27 33% (33)
want off + NP (Concrete)
North 41 80% (33)
Midland 95 ji93i%~_(i88)g,§‘ 12 .002
South 27 67% (18)
want off + NP (Abstract)
North 41 85% (3 5)
Midland 95 3719494 (89);; 1 8 .016
South 27 74% (20)

 

 

 

 

 

 

82

 

The hypothesis of a more widespread distribution of abstract senses of the forms of

N/W+PP and a more regional distribution of concrete senses, proposed in Section 2.4, is

not supported.

When concrete and abstract senses of the forms of N/W+PP are charted within

each region, as in Table 4.6, a different impression emerges about their use.

Table 4.6. Concrete and abstract senses of N/W+PP within each region

 

N

 

Concrete

 

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Abstract X: I (If I p

North

need in+NP 41/41 32% (13) 41%(17) 0.5 1 .492

need off+NP 41/41 39% (16) 51% (21) 0.8 1 .375

want in + NP 41/41 68% (28) 46% (19) 3.2 1 {.074ffﬁ

want off+ NP 41/41 80% (33) 85% (35) 0.1 1 .769
Midland

need in + NP 94/95 69% (65) 84% (80) 5.2 1 5.02351}-

need off+ NP 95/95 72% (68) 85% (81) 4.5 1 ;:.034fi1

want in + NP 95/95 91% (86) 80% (76) 3.4 1 .0651}:

want off + NP 95/95 93% (88) 94% (89) 0.0 1 1.0
South

need in + NP 27/27 22% (6) 48% (13) 2.9 1 5.087 ,

need off+ NP 27/27 26% (7) 52% (14) 2.8 1 1094 :

want in + NP 27/27 67% (18) 33% (9) 4.7 1 3029.3.

want off + NP 27/27 67% (18) 74% (20) 0.1 1 .766

 

 

 

 

 

 

 

In the North, the Midland, and the South, the use of abstract senses of need in + NP and

need oﬂ+ NP is greater than that of concrete senses; these differences are Si gniﬁcant in

the Midland (p=.023 and p=.034) and approach signiﬁcance in the South (p=.087 and

p=.094). The use of concrete and abstract senses of want oﬂ+ NP patterns like the forms

of N+PP in the North and the South: The use of abstract senses is greater than concrete

83

 

senses; however, these differences are not signiﬁcant. In each dialect region, want in +
NP shows a reversal of the trend: The use of concrete senses is greater than that of
abstract senses. These differences are Signiﬁcant in the South (p=.029) and approach
signiﬁcance in the Midland (p=.065) and in the North (p=.074).

The second hypothesis proposed in Section 2.4 that the use of abstract senses of
the linguistic variables is greater than concrete senses ﬁnds some support in the use of the
forms of N+PP. There are signiﬁcant or near signiﬁcant differences in the use of abstract
over concrete senses of need in + NP and need 017+ NP in the South and the Midland.
For the forms of W+PP, the hypothesis ﬁnds no support. For want 017+ NP, the
differences in the use of concrete and abstract senses are not signiﬁcant within each
region. Moreover, the differences in the use of concrete senses over abstract senses of
want in + NP are signiﬁcant or nearly signiﬁcant within each dialect region, contrary to

the hypothesis.

4.2.2. N+PA

The concrete and abstract senses of the forms of N+PA are presented in Table 4.7. As
with the regional distribution of the combined variables, seen in Table 4.2, the differences
in the use of concrete and abstract senses of the forms of N+PA across the regions are
signiﬁcant. For each form, use in the Midland is greater than in the North and the South.
The hypothesis of a more widespread distribution of the abstract senses of the forms of
N+PA and a more regional distribution of concrete senses, proposed in Section 2.4, is not

supported.

84

Table 47. Concrete and abstract senses of N+PA across dialect regions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pearson
N Use 12 I d f I p
need in (Concrete) -
North 41 59% (25)
Midland 95 vf89%;(184)jjf; 25 2 .000
South 27 52% (14)
need in (Abstract)
North 41 51% (21)
Midland 95 };_;8i‘21%(738)j5? 21 2 .000
South 27 44% (12)
need off (Concrete)
North 41 59% (24)
Midland3 49 f;90%(4;4);2; 14 2 .001
South 27 58% (15)
need off (Abstract)
North 41 37% (15)
Midland 95 “169%;(66);3_ig? 13 2 .001
South 27 52% (14)
need out (Concrete) .
North 41 41% (I7) .
Midland 94 :f80%(75)§i§§jig 27 2 .000
South 27 37% (10)
need out (Abstract)
North 41 54% (22)
Midland 94 77%(72) 7 2 .028
South 27 67% (18)

 

 

 

 

 

 

 

3 See footnote 1.

85

 

In Table 4.8, the usage statistics of concrete and abstract senses of the forms of

N+PA within each region are presented. A different pattern emerges here.

Table 4.8. Concrete and abstract senses of N+PA within each dialect region

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yates Corrected
N Concrete Abstract XI I d f j p

North

need in 41/41 59% (24) 51% (21) 0.2 .657

need off 41/41 59% (24) 37% (15) 3.1 {.077};

need out 41/41 41% (17) 54% (22) 0.8 .376
Midland

need in 95/95 90% (85) 82% (78) 1.6 .213

need off’ 49/95 90% (44) 69% (66) 6.3 301271;?

need out 94/94 80% (75) 77% (72) 0.1 .724
South

need in 27/27 52% (14) 44% (12) 0.1 .785

need off 26/27 58% (15) 52% (14) 0.0 .880

need out 27/27 37% (10) 67% (18) 3.6 3.057 g;

In the North, the Midland, and the South, the use of concrete senses of need in and need

oﬂ is greater than abstract senses. For need in, this difference is not signiﬁcant in any

dialect region. For need of, the difference in the use of concrete senses over abstract

senses is signiﬁcant in the Midland (p=.012) and approaches signiﬁcance in the North

(p=.077). The use of concrete senses and abstract senses of need out, however, patterns

differently within the dialect regions. In the North and South, the use of abstract senses of

need out is greater than that of concrete senses; in the South, this difference approaches

Signiﬁcance (p=.057); however, in the Midland the use of concrete senses is slightly,

though not signiﬁcantly, greater than that of abstract senses.

 

4 See footnote 1.

86

 

The second hypothesis proposed in Section 2.4 that the use of abstract senses of
the forms of N+PA is greater than that of concrete senses ﬁnds little support. Only in the
South does the difference in the use of abstract over concrete need out approach
Si gniﬁcance. Contrary to the hypothesis, differences in the use of abstract and concrete
senses of need in are not signiﬁcant in any dialect area, and the greater use of concrete
over abstract senses of need off in the North and Midland is signiﬁcant or approaches

Si gniﬁcance.

4.2.3. W+PA

The use of concrete and abstract senses of the forms of W+PA is presented in Table 4.9.
Interestingly, the use of concrete and abstract senses of the forms of W+PA patterns quite
differently from the overall results in Table 4.3, and from those of the forms of NIW+PP
and N+PA. The overall results for all forms of W+PA, shown in Table 4.3, as well as all
forms of N/W+PP and N+PA, including combined, concrete senses, and abstract senses,
shown in Tables 4.1, 4.2, 4.5, 4.7, all Show signiﬁcant differences in their distribution
across the dialect areas. For want in and want out, the differences in the use of concrete
senses across the dialect regions are signiﬁcant (p=.0419 and p=.OOO, respectively). The
Midland Shows much greater use of these forms (91%, 82%) than the North (78%, 51%)
and the South (70%, 41%). However, the differences in the use of abstract senses of want
in and want out are not Signiﬁcant; in fact, the use of abstract want in and abstract want
out in the North and South rivals that of the Midland. For want oﬂ, the pattern is
reversed: The use of concrete senses across the regions is not Si gniﬁcant; however, the

use of abstract senses is signiﬁcant (p=.009).

87

The hypothesis that abstract senses of the forms of W+PA are in more widespread

use, while concrete senses are more Midland, proposed in Section 2.4, is supported by the

data for want in and want out. The use of abstract want in and abstract want out is not

signiﬁcantly different across the dialect regions, while the use of concrete want in and

concrete want out is signiﬁcantly greater in the Midland than in the North and the South.

Table 4.9. Concrete and abstract senses of W+PA by region

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pearson
N Use x; T df I P
want in (Concrete)
North 40 78% (31) ::
Midland 95 1,29;1%T(86)§1§55; 8 2 .019, >
South 27 70%(19) ~ . - ~
want in (Abstract)
North 41 71% (29)
Midland 95 80% (76) 2 2 .434
South 27 81% (22)
want off (Concrete)
North 41 83% (34)
Midland 94 89% (84) 4 2 .128
South 27 74% (20)
want off (Abstract)
North 41 46% (19) 5,:
Midland5 49 ig78%i(38)§g{,§ 9 2 T009; '
South 27 63% (17) " '
want out (Concrete)
North 41 51% (21) ,»
Midland 95 :ii'8f2:%(7i8)§;;3j.; 23 2 .000.
South 27 41%(11) ; .
want out (Abstract)
North 41 ‘ 80% (33)
Midland 95 77% (73) 1 2 .623
South 27 85% (23)

 

 

 

 

 

 

 

5 See footnote 2.

88

 

The hypothesis is not supported and is turned on its head by the results of want 017': The
use of concrete senses of want of is more widespread, and the use of the abstract senses
is more regional. The use of concrete want oﬂ in the North, the Midland, and the South is
not signiﬁcantly different, while the use of abstract want of is signiﬁcantly greater in the
Midland than in the North and South.

The usage statistics of the forms of W+PA within each dialect region, shown in
Table 4.10, reveal some intriguing distributional differences of concrete and abstract
senses. Again, it appears that there is an interaction between the concrete and abstract

senses and the linguistic variables.

Table 4.10. Concrete and abstract senses of W+PA within each dialect region

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yates Corrected
N Concrete Abstract X1 I d f I p

North

want in 40/41 78% (31) 71% (29) 0.2 1 .659

want off 41/41 83% (34) 46% (19) 0.1 1 1.0017j 1

want out 41/41 51% (21) 80% (33) 6.6 1 .010],
Midland

want in 95/95 91% (86) 80% (76) 3.4 1 ’ .065

want oii6 94/49 89% (84) 78% (3 8) 2.7 1 .100

want out 95/95 82% (78) 77% (73) 0.5 l .472
South

want in 27/27 70% (19) 81% (22) 0.4 1 .524

want off 27/27 74% (20) 63% (17) 0.3 1 .558

want out 27/27 41% (11) 85% (23) 9.6 1 .002

 

 

 

 

 

 

 

The use of concrete want in is greater than abstract want in in the North and the Midland,

and in the Midland, this difference approaches Signiﬁcance (p=.065). In the South, this

 

6 See footnote 2.

89

 

pattern is reversed, but the difference between the use of concrete want in and abstract
want in is not signiﬁcant The use of concrete senses of want oﬂ is greater than that of
abstract senses in all three dialect areas. The difference in the use of concrete want of
over abstract want off is signiﬁcant in the North (p=.001) and approaches signiﬁcance in
the Midland (p=. 100). The use of abstract senses of want out is greater than that of
concrete senses in the North and South, and these differences are signiﬁcant (p=.010 and
p=.002, respectively). In the Midland, the use of abstract want out is lower than that of
concrete want out; however, this difference is not signiﬁcant. Some parallels emerge
between concrete and abstract senses of W+PA and N+PA. The distribution of the use of
concrete senses and abstract senses of want of parallels that of need oﬁi while the
distribution of concrete senses and abstract senses of want out parallels that of need out.
The second hypothesis proposed in Section 2.4 that abstract senses of W+PA are
used more than concrete senses is supported by the distribution of want out; however, the

hypothesis is not supported by the data for want in and want off

4.2.4. Discussion

The semantic features of concrete and abstract appear to play a role in the use of N/W+PP
and NIW+PA, though this role is not as clear cut as one might wish. Across the three
dialect areas, the use of both concrete senses and abstract senses of the forms of NIW+PP
and N+PA is Signiﬁcantly greater in the Midland than in the North and the South,
contrary to the hypothesis that only concrete senses are used more in the Midland, while
abstract senses are more widespread. For W+PA, the pattern is more complex: Midland

use is signiﬁcantly greater than Northern use and Southern use for concrete senses of

90

want in and want out (but not for abstract senses of want in and want out) and for abstract
senses of want oﬂ (but not for concrete senses of want oﬁ). The data from want in and
want out support the hypothesis, while those from want oﬂ do not.

Why Should only the forms of W+PA be distributed differently across the
regions? Perhaps the reason lies in the overall distributions seen in Section 4.1. AS
evident in Table 4.5, want in, want out, and want oﬁ" are the only forms of NIW+PP and
N/W+PA for which the use by respondents in all three dialect areas is close in range, i.e.
less than 20%. Thus, the fact that use of these forms in the North and the South is more
comparable to use in the Midland allows for nonsigniflcant differences to emerge. For all
of the other forms, use in the Midland is so much greater than in the North and South that
signiﬁcant differences inevitably result. I do not have an explanation as to why concrete
and abstract senses of want in and want out pattern differently from those of want oﬂ

Furthermore, the forms of NIW+PP and NIW+PA are treated remarkably
Similarly by respondents within each dialect region. Three tendencies in the use of
concrete and abstract senses of the forms of N/W+PP and NIW+PA seem apparent, and
can be seen in Table 4.11: 1) Abstract senses of need in + NP, need off+ NP are used
more than concrete senses by respondents in all three dialect areas; this trend also holds
for need out and want out in the North and the South. In Table 4.11, these items are
Shaded in dark gray and tend to be clustered more at the lesser end of the usage scale,
especially in the North and the South. 2) Concrete and abstract senses of want off+ NP
are used more or less equally by respondents in all three dialect areas, as is the case for
need in and want in in the North and the South. In Table 4.11, these items are unshaded

and seem to cluster more at the greater end of the usage scale. 3) Concrete senses of want

91

in + NP are used more than abstract senses by respondents in all three dialect areas, as is
the case for need 01% and want oﬂ in the North and Midland. In Table 4.11, these items
are shaded in light gray and tend to be scattered throughout the usage scale, not clustered

at one end.

Table 4.11. Rank order of concrete and abstract senses of N/W+PP and NIW+PA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

North Use Midland Use South Use
_ineed 1n + NP (con) 32% need off (abs) 69% ineed 1n + NP (con) 22%
need off (abs) 37% :need in + NP (con) 70% :need off + 531*}..3' (con) 26%
need off + NP (con) 39% éneed off + NP (con) 2 72% Want in + NP (abs) 33%
ineed 1n + NP (abs) 41% need out (abs) 77% need out (con) """" 37%
need out (con) 141% want out (abS) , 77% Want out (con) ,,,,,,,,,,, 41%
want in + NP (abs) . 46% want off (abs) 3 ‘ 78% need in (abs) 44%
want off (abs) 33 Z16% Want in + NP (abs) 80% need in + NP (abs) 48%
heed off + NP (abs) 1 51% want in (abs) _ , 80% need off + NP (abs) 3 52%
need 1n (abs) 51% need out (con) 80% need in (con) 52%
Want out (con) 511% want out (con) ’ . ' - 82% need off (abs) 52%
need out (abs) 54% need in (abs) 82% need off (con) 58%
need in (con) 59% need in + NP (abs) 84% want off (abs) 63%
need Off (con) ; * J 59% needoff +NP(abs) 85% want in + NP (con) 67%
want in + NP (con) ’3 68% want off (con) 1 ’ 89% want off + NP (con) 67%
want in (abs) 71% needin (con) » . *3 89% need out(abs) 67%
want in (con) 78% need off (con) * 90% want in (con) 70%
want off + NP (con) 80% want in + NP (con) 91% want off (con) 70%
Want out (abs) 80% want in (con) j . 91% want off + NP (abs) 74%
want off (con) .1 . , 83% want off + NP (con) 93% want in (abs) 81%
want off + NP (abs) 85% want off + NP (abs) 94% wantouttabs) 85%

 

Though not entirely conclusive, these trends raise questions about the role of
concrete and abstract features in language change. The tendency in semantic change is for

abstract senses to develop out of concrete senses (e.g. Williams 1975:207; Traugott 1982,

92

1995; Langacker 1999; Liidtke 1999). The data presented here seem to suggest that
abstract senses ﬁrst take hold, as evidenced by the greater use of abstract senses over
concrete senses of forms that are less used in the hierarchy, those Shaded in dark gray in
Table 4.11, and with increased use, the concrete and abstract senses level out.
Admittedly, this is Speculative, and the small sample sizes, particularly in the North and
the South, caution against overinterpreting the data. More research is needed with larger
numbers of Speakers in communities in all three dialect areas, but particularly in the
North and South, to substantiate these ﬁndings.

In summary, an examination of semantic distinctions like the differences in use
between concrete and abstract senses of N/W+PP and NIW+PA reveals some intriguing
patterns across and within the dialect areas: 1) There are few differences in the
distribution of concrete and abstract forms across dialect areas, with the exception of
W+PA forms. The hypothesis that concrete senses Show a more Midland distribution
while abstract senses Show more widespread distribution is not supported for the forms of
N/W+PP, N+PA, and want 01% on the other hand, the hypothesis is supported for the
forms of want in and want out. 2) There are some notable differences in the treatment of
concrete and abstract senses within the dialect areas; in some cases, abstract senses are
used more than concrete senses, e.g. need in + NP and need oﬂ+ NP; in others, concrete
senses are used more than abstract senses, e. g. want in + NP; and in others, concrete and
abstract senses are used nearly equally, e. g. want off+ NP. 3) Individual forms of
N/W+PP and N/W+PA behave quite similarly within the dialect areas; for example, the
forms cited in 2 followed the respective patterns in all three dialect areas. There seem to

be implications here for the ﬁeld of semantic change; however, additional research with

93

greater numbers of respondents, particularly in the North and the South, is needed to

enable conclusions about language change.

94

4.3. Usage hierarchies

In the previous sections, some usage hierarchies of individual prepositional phrases and
prepositional adverbs as well as concrete senses versus abstract senses were presented.
This section examines usage hierarchies between the use of need and want in PP and PA

constructions and between the use of PP and PA constructions.

4.3.1 Need vs. want

In the North, the Midland, and the South, there is a clear preference for want over need
both in prepositional phrase constructions and in prepositional adverb constructions.
Figure 4.3 displays an overall comparison of the use of need and want in PP

constructions.

100% 1

89V
90% — °

78%

80% 7

70%

70%
60%
60%
50%

40% I

37%

30%
20%

  

10%

0% . . . . . .

North (N=164) Midland (N=380) South (N=108)
[ raN+PP l:lW+PP I

Figure 4.3. Use of N+PP vs. W+PP in the North (x2 =27, df 1, p=.000), the Midland
(x2 =18, df 1, p=.000), and the South(x2=ll,df1, p=.001)

 

 

 

 

 

 

“w. .~.s \

 

 

95

The differences between the use of N+PP and W+PP are signiﬁcant in the North
(p=.000), the Midland (p=.000), and the South (p=.001). The overall data from the North,
the Midland, and the South provide strong support for the hypothesis that the use of
W+PP is greater than the use of N+PP.

The use of the forms of N+PP versus W+PP within each region are shown in
Table 4.12. In every case, the use of want is greater than need. The differences in the use
of need in/0ff+ NP and want in/0ﬁ+ NP are signiﬁcant in the North (p=.012, p=.000)
and the Midland (p=.047, p=.000). In the South, however, only the differences between
the use of need oﬂ+ NP and want oﬂ+ NP are signiﬁcant (p=.002). In the North and the
Midland, the data provide strong support for the hypothesis that want is more acceptable
than need in prepositional phrase constructions; in the South, the hypothesis holds only

for 0ﬂ+ NP.

Table 4.12. Use of N+PP vs. W+PP

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yates Corrected
N Need Want X2 I df L p
_ North ‘
in + NP 82/82 37% (30) 57% (47) 6 1 .012
off + NP 82/82 45% (37) 83% (68) 24 .000
Midland
in + NP 189/190 77%(145) 85% (162) 4 .047
off+ NP 190/190 78% (149) 93% (177) 16 .000
South
in + NP 54/54 35%(19) 50% (27) 2 .173
off + NP 54/54 39% (21) 70% (38) 10 .002

 

96

 

In all three dialect areas, there appears to be an interaction between the use of

need and want and the use of concrete and abstract senses. Table 4.13 presents the usage

statistics for the use of concrete and abstract senses of N+PP vs. W+PP.

Table 4.13. Use of concrete and abstract senses of N+PP vs. W+PP

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yates Corrected
N Need Want x2 I d f J p
North
in + NP (Concrete) 41/41 32% (13) 68% (28) 9.6 1 13.002}
in + NP (Abstract) 41/41 41% (17) 46% (19) 0.0 1 .824
off+ NP (Concrete) 41/41 39% (16) 80% (33) 13.0 1 .27.:000177
off+ NP (Abstract) 41/41 51% (21) 85% (35) 9.5 1 {002]
Midland
in + NP (Concrete) 94/95 69% (65) 91% (86) 12.1 1 1.00071
in + NP (Abstract) 95/95 84% (80) 80% (76) 0.3 1 .570
off + NP (Concrete) 95/ 95 72% (68) 93% (88) 12.9 1 [.0003
off+ NP (Abstract) 95/95 85% (81) 94% (89) 2.7 1 2.098 :
South
in + NP (Concrete) 27/27 22% (6) 67% (18) 9.1 1 ‘ 903?. 3*
in + NP (Abstract) 27/27 48% (13) 33% (9) 0.7 1 .406
off+ NP (Concrete) 27/27 26% (7) 67% (18) 7.4 1 .0065
l off+ NP (Abstract) 27/27 52% (14) 74% (20) 2.0 1 .159

 

In all three dialect areas, the differences between the use of concrete need in/off+ NP and

concrete want in/oﬂ+ NP are Signiﬁcant. The differences in the use of abstract need in/oﬂ

+ NP and abstract want in/oﬂ+ NP tend not to be signiﬁcant. Only in the North is the

difference between the use of abstract need Off+ NP and abstract want oﬂ+ NP

Si gniﬁcant (p=.002), though in the Midland the difference approaches signiﬁcance

(p=.098). In general, concrete senses of the PP constructions support the hypothesis that

97

 

the use of want is more acceptable than the use of need, while abstract senses Show more
equal use of need and want.

Figure 4.4 displays an overall comparison of the use of need and want in PA
constructions. The difference in the use of N+PA and W+PA is signiﬁcant only in the
North (p=.000) and the South (p=.000). Only in the North and the South do the data

support the hypothesis that W+PA is used more than N+PA.

 
       
 
   
  

   
 

 

 

 

 

 

 

 

 

 

 

100% I

90% ‘ 81% 84%

70% I 68% ,1 69 A.
60%

50%

40%

30%

20%

10% g, _:. _3

North (N=328) Midland (N=713) South (N=2l6)

 

1 IN+PA l:lW+PA I

Figure 4.4. Use of N+PA vs. W+PA in the North (x2 =30, df 1, p=.000), the Midland (x2
=2, df 1, p=.l43), and the South(x’=14, df 1, p=.001)

 

Table 4.14 shows the usage statistics for the forms of N+PA and W+PA. The
differences in the use of N+PA and W+PA are not as clear cut as those of N+PP and

W+PP.

98

Table 4.14. Use of N+PA vs. W+PA

 

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N NCCd Want x1 I (If I p
North _
by 41/41 34%(14) 51%(21) 1.8 l .180
down 41/41 39% (16) 85% (35) 17.8 1 .7000 f
in 82/81 55% (45) 74% (60) 5.7 1 .016
off 82/82 48% (39) 65% (53) 4.2 l .041
out 82/82 48% (39) 66% (54) 4.9 1 .027 v
Midland
by 95/95 80% (76) 77% (73) 0.1 l .724
down 95/95 82% (78) 94% (89) 5.0 1 2.026
in 190/190 86% (163) 85% (162) 0.0 l 1.0
off7 144/143 76% (110) 85% (122) 3.1 l - .077
out 188/190 78% (147) 79% (151) 0.0 1 .858
South
by 27/27 48% (13) 52% (14) 0.0 1 1.0
down 27/27 44% (12) 81% (22) 6.4 1 - .011
in 54/54 48% (26) 76% (41) 7.7 l 006
off 53/54 55% (29) 69% (37) 1.6 1 .204
out 54/54 52% (28) 63% (34) 0.9 1 .331

 

 

 

 

 

 

 

Especially in the North, and, to some extent in the South, there seems to be a preference

for the forms of W+PA over N+PA: In general, the use of forms of W+PA is greater than

the use of corresponding forms of N+PA. In the North, these differences are signiﬁcant

for the prepositional adverbs down (p=.000), in (p=.016), oﬂ (p=.041), and out (p=.027).

In the South, these differences are signiﬁcant for the prepositional adverbs down(p=.011)

and in (p=.006). In the Midland, the use of need and want with the prepositional adverbs

by, in, oﬂ and out is very Similar, and these differences are not signiﬁcant, though for oﬂ,

the differences approach Si gniﬁcance (p=.077). In the Midland, the preference of want

 

7 See footnotes l and 2.

99

 

over need is Significant, however, for the prepositional adverb down (p=.026). In fact,
only need down and want down Show signiﬁcant differences in all three dialect regions.
This may be due to the fact that want down is the most used form in all three dialect
regions, as well (see Section 4.1.3). The hypothesis that the prepositional adverb
constructions with want will be used more than those with need ﬁnds quite a bit of
support in the North, less support in the South, and little support in the Midland.

Going to the level of detail of concrete and abstract senses of the forms of N+PA
and W+PA reveals little more of interest, as shown in Table 4.15. In the North and South,
the trend for greater use of want than need in the forms of the PA constructions remains
consistent; however, few forms show signiﬁcant differences in use: In the North, the
differences in the use of need and want with concrete of and abstract out are signiﬁcant
(p=.029, p=.019), and in the South only abstract in Shows signiﬁcance (p=.011). In the
Midland, there are no Signiﬁcant differences between the use of N+PA and W+PA in

concrete and abstract senses.

100

Table 4.15. Use of concrete and abstract senses of N+PA vs. W+PA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Yates Corrected
N Need Want x7 I df—I P
North .
in (Concrete) 41/40 59% (24) 78% (31) 2.5 1 .112
off (Concrete) 41/41 59% (24) 83% (34) 4.8 1 7.029 3
out (Concrete) 41/41 41% (17) 51% (21) 0.4 1 .506
in (Abstract) 41/41 51% (21) 71% (29) 2.5 1 .113
off (Abstract) 41/41 37% (15) 46% (19) 0.5 1 .501
out (Abstract) 41/41 54% (22) 80% (33) 5.5 1 3.019 '
Midland
in (Concrete) 95/95 89% (85) 91% (86) 0.0 1 1.0
off (Concrete)8 49/95 90% (44) 89% (84) 0.0 1 1.0
out (Concrete) 94/95 80% (75) 82% (78) 0.0 1 .825
in (Abstract) 95/95 82% (78) 80% (76) 0.0 1 .853
off (Abstract)’ 95/49 69% (66) 78% (3 8) 0.7 1 .407
out (Abstract) 94/95 77% (72) 77% (73) 0.0 1 1.0
South
in (Concrete) 27/27 52% (14) 70% (19) 1.2 l .264
off (Concrete) 26/27 58% (15) 74% (20) 0.9 l .333
out (Concrete) 27/27 37% (10) 41% (11) 0.0 1 1.0
in (Abstract) 27/27 44% (12) 81% (22) 6.4 1 ;.011
off (Abstract) 27/27 52% (14) 63% (17) 0.3 1 .582
out (Abstract) 27/27 67% (18) 85% (23) 1.6 1 .203

 

 

 

 

 

 

 

4.3.2. NIW+PP vs. N/W+PA

The overall use of PP versus PA constructions is presented in Table 4.16. In the North,

 

the Midland, and the South, the overall use of N+PA is greater than the use of N+PP. In
the South, this difference is Signiﬁcant (p=.034). With respect to the use of want, only in
the South is the use of the PA construction greater than the PP construction. In the North

and Midland, the use of W+PP is greater than that of W+PA, and in the Midland this

 

8 See footnote 1.
9 See footnore 2.

101

difference is signiﬁcant (p=.018). The hypothesis that the PA constructions are more

acceptable than the PP constructions is not supported by the overall data.

Table 4.16. General use of PP vs. PA

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N PP PA Yates Corrected
7:2 df P

North

need 164/328 41% (67) 47% (153) 1.3 1 .262

want 164/327 70% (115) 68% (223) 0.1 1 .740
Midland

need 379/712 78% (294) 81% (574) 1.2 1 .267

want 380/713 89% (339) 84% (597) 5.6 l i .018 {3
South

need 108/215 37% (40) 50% (108) 4.5 1 1.034

want 108/216 60% (65) 69% (148) 1.9 l .172

 

 

 

 

 

 

 

Table 4.17 presents the usage statistics for the forms of N/W+PP versus
N/W+PA. In all three dialect areas, the use of the PA construction of need in is greater
than the PP construction; these differences are signiﬁcant in the North (p=.028) and the
Midland (p=.033). In the North and the South, the use of the PA construction of want in
is signiﬁcantly greater than the PP construction (p=.013 and p=.010, respectively). In the
South, need off also Shows greater PA use than PP use, though this difference is not
Si gniﬁcant. In the North and Midland, want oﬂ reverses the trend and shows signiﬁcantly
greater use of the PP construction than the PA construction (p=.013 and p=.027,

respectively).

102

 

Table 4.17. Use ofPP vs. PA

 

 

 

PP

 

PA

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7:2 df P

North

need in 82/82 37% (30) 55% (45) 4.8 1 5.0285"?

need off 82/82 45% (37) 48% (39) 0.0 1 .876

want in 82/81 57% (47) 74% (60) 9.9 1 4.037;:

want off 82/82 83% (68) 65% (53) 6.2 1 1,01}
Midland

need in 189/190 77% (145) 86% (163) 4.5 1 11";033-3111'

need offlo 190/144 78% (149) 76% (110) 0.1 1 .758

want in 190/190 85% (162) 85% (162) 0.0 1 1.0

want off“ 190/143 93% (177) 85% (122) 4.7 1 1.02722
South

need in 54/54 35% (19) 48% (26) 1.4 1 .241

need off 54/53 39% (21) 155% (29) 2.1 1 .148

want in 54/54 50% (27) 76% (41) 6.7 1 "£2,010;

want off 54/54 70% (3 8) 69% (37) 0.0 1 1.0

 

 

 

 

 

 

 

4.3.5. Discussion

In general, in all three dialect areas, there is a preference for the use of W+PP over N+PP,

in support of the hypothesis. Concrete senses of the PP constructions tend to exhibit this

trend more strongly than abstract senses. The trend of greater use of want over need holds

for PA constructions, especially in the North and the South, and less so in the Midland. If

direct mention of a given construction in the previous literature is any indication of its

actual use, then it is not surprising that the use of the want constructions is much greater

than the use of the need constructions, since there are myriad observations and citations

of W+PA, and to a lesser extent W+PP, and few if any of N+PA and N+PP in the

 

10 See footnote 1.
11 See footnote 2.

103

 

literature. Furthermore, these ﬁndings hint at a possible hierarchy in language use and
change. In terms of use, it may be the case that in order to use need forms, one must also
have the corresponding want forms. In terms of language change, if change is indeed at
work in the North and the South, these ﬁndings Show that want forms are leading need
forms. Additional research is needed, particularly in the North and South.

Overall, with respect to the use of all PP and PA constructions with need, N+PA
constructions are used more than N+PP constructions, and this difference is Signiﬁcant in
the South. The use of all PP and PA constructions with want differs in the three dialect
areas: In the North and the Midland, W+PP constructions are used more than W+PA, and
this difference is signiﬁcant in the Midland, while W+PA constructions are used more
than W+PP in the South. At the level of the individual form, in the North and the South,
the individual forms need in/oﬂ and want in tend to be used more in the PA construction
than the PP construction, the exception is want of, which is used more in the PP
construction than the PA construction. In the Midland, the PA construction of need in is
used more than the PP construction; however, the PA and PP constructions of need of
and want in are used nearly equally, and the PP constructions of want of are used more
than the PA constructions. Interestingly, in all three dialect areas, want of is used more or
about the same in PP constructions than PA constructions. Recall that want 017+ NP and
want oﬂ were, inexplicably, among the most used forhrs of NIW+PP and NIW+PA, as
shown in Table 4.4. Otherwise, it would appear that the less used a form is, the more

likely it is to be of PP constructions.

104

4.4. Nonlinguistic factors

The distribution of NIW+PP and NIW+PA is examined with respect to several
nonlinguistic variables: rural-urban residence (Section 4.4.1), gender (Section 4.4.2),
socioeconomic status (Section 4.4.3), and age (Section 4.4.4). In Section 4.4.1 and
Section 4.4.2 the use of N/W+PP and N/W+PA across the nonlinguistic variables of
rural-urban residence and gender are examined in the North, the Midland, and the South.
In Section 4.4.3 and Section 4.4.4 the use of the linguistic variables across the
nonlinguistic variables of socioeconomic status and age are examined only in the
Midland. The small samples sizes in the North and South do not permit statistically

reliable analyses across these nonlinguistic variables With multiple groupings.

4.4.1. Rural-urban

Tables 4.18-4.20 present the usage statistics for the forms of N/W+PP and N/W+PA by
respondents from rural and urban areas within each dialect region. For every form of
NIW+PP and N/W+PA, the differences in their use by rural and urban dwellers are not
Signiﬁcant nor do any of them approach signiﬁcance. These ﬁndings support the
hypothesis, prOposed in Section 2.5.1, that there are no signiﬁcant differences in the use
of the forms of NIW+PP and NIW+PA by rural and urban dwellers within each dialect

area.

105

Table 4.18. Use of NIW+PP by rural-urban residence

 

 

 

 

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N Rural Urban XI d f p

North ’

need in + NP 16/66 44% (7) 35% (23) 0.1 1 .708

need off + NP 16/66 44% (7) 45% (30) 0.0 l 1.0

want in + NP 16/66 56% (9) 58% (38) 0.0 1 1.0

want off+ NP 16/66 94% (15) 80% (53) 0.8 1 .362
Midland

need in + NP 55/134 76% (42) 77% (103) 0.0 1 1.0

need off+ NP 56/134 71% (40) 81% (109) 1.7 1 .186

want in + NP 56/134 80% (45) 87% (117) 1.0 1 .313

want off + NP 56/134 95% (53) 93% (124) 0.4 l .834
South

need in + NP 12/42 42% (5) 33% (14) 0.04 1 .849

need off+ NP 12/42 33% (4) 40% (17) 0.01 1 .911

want in + NP 12/42 42% (5) 52% (22) 0.11 1 .743

want off + NP 12/42 67% (8) 71% (30) 0.00 1 1.0

 

 

 

 

 

 

 

106

 

Table 4.19. Use of N+PA by rural-urban residence

 

 

N

 

Rural

 

Urban

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

7:2 df 1 P
North
need by 8/33 25% (2) 36% (12) .04 1 .847
need down 8/33 50% (4) 36% (12) .09 1 .760
need in 16/66 50% (8) 56% (37) .02 l .875
need off 16/66 44% (7) 48% (32) .00 1 .951
need out 16/66 50% (8) 47% (31) .00 1 1.0
Midland
need by 28/67 79% (22) 81% (54) .00 l 1.0
need down 28/67 82% (23) 82% (55) .00 1 1.0
need in 56/ 134 80% (45) 88% (118) 1.34 1 .247
need off” 42/102 79% (33) 75% (77) .03 1 .857
need out 56/132 70% (39) 82% (108) 2.74 1 .098
South
need by 6/21 67% (4) 43% (9) .32 1 .571
need down 6/21 67% (4) 38% (8) .44 1 .438
need in 12/42 42% (5) 50% (21) .03 l .856
need off 12/41 42% (5) 59% (24) .49 1 .482
need out 12/42 50% (6) 52% (22) . .00 1 1.0

 

 

 

 

 

 

 

 

'2 See footnote 1.

107

 

Table 4.20. Use of W+PA by rural-urban residence

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N Rural Urban Yates Corrected
1’ df P
North
want by 8/33 50% (4) 52% (17) .00 l 1.0
want down 8/33 100% (8) 82% (27) .55 1 .455
want in 16/65 69% (11) 75% (49) .05 1 .823
want off 16/66 81% (13) 61% (40) 1.58 1 .208
want out 16/66 63% (10) 67% (44) .00 1 .983
Midland
want by 28/67 71% (20) 79% (53) .29 1 .588
want down 28/67 96% (27) 93% (62) .06 1 .804
want in 56/134 91% (51) 83% (111) 1.53 1 .217
want off13 41/102 85% (35) 85% (87) .00 1 1.0
want out 56/134 75% (42) 81% (109) .62 1 .430
South
want by 6/21 67% (4) 48% (IO) .13 I .719
want down 6/21 67% (4) 86% ( 18) .21 1 .643
want in 12/42 83% (10) 74% (31) .09 l .766
want off 12/42 75% (9) 67% (28) .04 1 .845
want out 12/42 50% (6) 67% (28) .51 1 .474
4.4.2. Gender

Table 4.21 presents the distribution of the forms of N/W+PP by men and women. Most

 

notable are the signiﬁcant differences in the use of these forms between men and women
in the South. In the South, men’s use over women’s use is signiﬁcant or approaches

Si gniﬁcance for need in +NP (p=.010), want in + NP (p=.007), and need 017+ NP
(p=097). In the North, men use want oﬂ+ NP Signiﬁcantly more than women (p=.035).

In the Midland, use of these forms is similar for men and women, though in the case of

 

’3 See footnote 2.

108

need in + NP, the greater use by women than men approaches signiﬁcance (p=.084). The

hypothesis, proposed in Section 2.5.4, that there are no differences in the use of NIW+PP

between men and women ﬁnds mixed support. In the Midland and the North, at least

three of the four forms support the hypothesis; however, in the South three of the four

forms do not support the hypothesis.

Table 4.21. Use of NIW+PP by men and women

 

 

 

 

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N Men Women x2 df p

North

need in + NP 62/20 37% (23) 35% (7) .00 1 1.0

need off+ NP 62/20 48% (30) 35% (7) .62 1 .431

want in + NP 62/20 58% (36) 55% (11) .00 1 1.0

want off+ NP 62/20 89% (55) 65% (13) 4.44 1 ;.035{’
Midland

need in + NP 128/61 73% (93) 85% (52) 3.00 1 5.084 3 f

need off+ NP 128/62 78% (100) 79% (49) * .00 1 1.0

want in + NP 128/62 85% (109) 85% (53) .00 1 1.0

want off+ NP 128/62 95% (121) 90% (56) .59 1 .441
South

need in+NP 38/16 47% (18) 6% (1) 6.64 1 5.010.

need off+ NP 38/16 47% (18) 19% (3) 2.77 1 f.096 . 1

want in + NP 38/16 63% (24) 19% (3) 7.19 1 5.007 .:

want off+NP 38/16 71% (27) 69% (11) .00 1 1.0

 

 

 

 

 

 

 

Table 4.22 displays the results of the use of N+PA by men and women. Similar to

the pattern of distribution of N/W+PP, in the North and South, men use the forms of

N+PA more than women, with one exception, need out in the North. There are signiﬁcant

or nearly signiﬁcant differences between men’s and women’s use of need in in the North

(p=.072) and in the South (p=.002) and need on! in the South (p=.095). In the Midland,

109

 

the use by men and women, in general, is similar and is not signiﬁcant, with one

exception: The use of need out by men is signiﬁcantly greater than by women (p=.040).

Table 4.22. Use of N+PA by men and women

 

 

 

 

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

N Men Women x; I d f I p
North
need by 31/10 39% (12) 20% (2) .49 1 .483
need down 31/10 42% (13) 30% (3) .09 1 .764
need in 62/20 61% (3 8) 35% (7) 3.23 1 ~(0723;?
need off 62/20 47% (29) 50% (10) .00 1 1.0
need out 62/20 45% (28) 55% (11) .26 1 .611
Midland
need by 64/31 80% (51) 81% (25) .00 1 1.0
need down 64/31 81% (52) 84% (26) .00 1 .978
need in 128/62 86% (110) 85% (53) .00 1 1.0
need off” 92/52 76% (70) 77% (40) .00 1 1.0
need out 128/60 83% (106) 68% (41) 4.21 1 f.040f
South
need by 19/8 58% (11) 25% (2) 1.30 1 .254
need down 19/8 47% (9) 38% (3) .00 1 .962
need in 38/16 63% (24) 13% (2) 9.63 1 {.0072}
need off 37/16 59% (22) 44% (7) .57 1 .451
need out 38/16 61% (23) 31% (5) 2.78 1 5.095”-

 

 

 

 

 

 

 

Table 4.23 presents the results of the use of W+PA by men and women. In the

North and South, men tend to use the forms of W+PA more than women, with a few

exceptions. In the South, the greater use by men than by women of want of is signiﬁcant

(p=.026) and want in is nearly Signiﬁcant (p=.065). In the Midland, for want down and

 

’4 See footnote 1.

110

 

want of, use by women exceeds use by men; however, for want by, want in, and want out

use by men is greater than use by women. For want by this difference is nearly

signiﬁcant (p=.085).

Table 4.23. Use of W+PA by men and women

 

 

N

 

Men

 

Women

Yates Corrected

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

x’ 1 df l P
North
want by 31/10 58% (18) 30% (3) 1.39 1 .238
want down 31/10 90% (28) 70% (7) 1.14 1 .286
want in 61/20 77% (47) 65% (13) .60 1 .439
want off 62/20 63% (39) 70% (14) .10 1 .758
want out 62/20 66% (41) 65% (13) .00 1 1.0
Midland
want by 64/31 83% (53) 65% (20) 2.97 1 . .085 1‘
want down 64/31 92% (59) 97% (30) .17 1 .680
want in 128/62 87% (111) 82% (51) .35 1 .552
want off15 92/51 84% (77) 88% (45) V .24 1 .626
want out 128/60 82% (105) 74% (46) 1.13 1 .288
South
want by 19/8 63% (12) 25% (2) 1.93 1 .164
want down 19/8 89% (17) 63% (5) 1.22 1 .269
want in 38/16 84% (32) 56% (9) 3.41 1 1.065
want off 38/16 79% (30) 44% (7) 4.94 1 .026 A ’
want out 38/16 55%(21) 81% (13) 2.24 1 .134

 

 

 

 

 

 

 

The hypothesis, proposed in Section 2.5.4, that there are no differences among men and

women in their use of NIW+PA is supported for many forms in the Midland and the

North but for few forms in the South. Discussion of these patterns follows in Section

4.4.5.

 

’5 See footnote 2.

111

 

4.4.3. Socioeconomic status

Table 4.24 presents the socioeconomic distribution of NIW+PP in the Midland across
ﬁve socioeconomic classes: upper middle, middle middle, lower middle, upper working,
and lower working. There are no signiﬁcant differences in the use of the forms of
N/W+PP across socioeconomic classes. In fact, no socioeconomic class dominates in the
use of any of the forms of N/W+PP.

One interesting observation, however, is the equal use of need in + NP and want
in + NP among respondents in the lower middle and lower working classes. In the other
socioeconomic classes, want in + NP is used more than need in + NP, and overall in the
Midland the greater use of want in + NP over need in + NP was found to be signiﬁcant
(see Section 4.2). Nevertheless, the hypothesis that there are no signiﬁcant differences in
the distribution of N/W+PP across socioeconomic classes in the Midland, proposed in
Section 2.5.2, is supported by the data.

Table 4.25 presents the results of the use of N/W+PA across the socioeconomic
classes. The only variable whose distribution is Si gniﬁcant across the socioeconomic
classes is want out (p=.002). Interestingly, the use of want out is greater among the lower
middle class (94%) than among the upper working and lower working class (62% and
68%, respectively). The hypothesis that there are no Signiﬁcant differences in the use of
the forms of N/W+PA across the socioeconomic classes in the Midland, proposed in

Section 2.5.2, ﬁnds ample support in the data.

112

.N 08508 new 2
._ 08508 com o.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

So. a 8.2 a a same as .3% as sad as :3 69 so” Shamans? :5 ans
2 _. e 42. 8: 3% AB on; $9 33.3 a: $8 ARV axes M2: Q9225. :mo :33
was. a on... SC on: a a :8 5e sea as $3 $9 exam Shamans? 5 are
an a 23 av saw as as: as sea _ a: :8 as name _SQeQE—N Econ ans
mom. a a? a: :5 6: sea. a: ﬁne 2 a $2 6: son :zsesezs 3 ans
Em. a mg as sac as $9 AB on: as $3 9.3 $3 «gangs? so coon
com. a mam Q: :2 :8 can ecu see 9.: as: as each @3332 2&6 bone
ham. 6 8a a: .38 A5 5” ci £8 £3 :3 $9 $3 8?ng? a one:
was. e Ed Q .3% :5 as; as $3 2 a $2 Ev as; 5554:: :38 been
new. a ”3 8: as; 9: as; a: «an Ag 3% Ev as; 55.53% .3 been
a =. ax wafers ”53:63 22:2 22:2 22::
59.3.— ..oBeA .895 33.5 22:2 .895 Z 1:232
320 28058208 3 <m+><Z mo 83 and 2an
com. 4 own a: so” a a :8 63 goo 63 x8 8"; $8 a2 néo Bea
Sm. e m? a: $8 63 New 83 £3 as $8 A5 saw dz + 5 ans
5.. e Nam 93 as? as see as $3 :3 as? as as: m2 lac one:
New. a 5% a: snow as as: as; 3% a: sec $3 $2 m2 + a one:
a E ax w Suz :ttz %uz anuz Neuz
€263 ”£3.33 22:2 6.632 23.2
:omaaom auto: 5233 .5364 9.352 .395 .5222

 

 

 

 

 

 

 

 

mam—o oaocooooEOm r3 mm+2<z mo 33 .34 03mm.

113

Some other interesting observations can be found in the data in Table 4.25. First,
the lower middle class is among the greatest users of these forms across all
socioeconomic classes; in particular, the lower middle class leads in the use of need in
and in the use of all forms of W+PA, with the exception of want by. Second, the upper
middle class is never the greatest or least users of any of the forms of NIW+PA. Reasons

for these observations are explored in more detail in Section 4.4.5.

4.4.4. Age

Table 4.26 presents the distribution of NIW+PP across four age groups in the Midland.
The differences in the use of need in + NP and want in + NP among the age groups
approach Signiﬁcance (p=.057 and p=.098, respectively); however, the differences in the
use of need oﬂ+ NP and want oﬂ+ NP among the age groups are not signiﬁcant. The

Si gniﬁcance in the age distribution of need in + NP and want in + NP is likely due to the
lesser rate of use of these forms among the 55+ years group. In fact, the 55+ years group
uses three of the four variables—need in + NP, need oﬂ+ NP, and want in + NP—the
least among all of the age groups. Interestingly, the 55+ years group has a greater rate of
use of want oﬂ+ NP, comparable to that of the other age groups. The hypothesis that
there are no signiﬁcant differences in the use of N/W+PP among respondents of different

age groups in the Midland, proposed in Section 2.5.3, is not supported by the data.

114

Table 4.26. Use of N/W+PP by age

 

 

 

 

 

 

 

 

 

 

 

 

 

. - - 55+ rs Pearson
M1 dlan d 1815:2531 25N3=42§rs 351$;er N=2yO x2 df F
need in+NP 80%(103) 73%(16) 80%(16) H53%(10)fji 7.5 3 3.057
need off+NP 80%(103) 77%(17) 80%(16) 5,65%(13‘); 2.5 3 .477
wantin+NP 89%(114) 77%(17) 85%(17) ju'70%1(14)1j6.3 3 .093,"
want off+NP 94%(120) 86%(19) 95%(19) 95%(19) 19 3 .598

 

 

 

The usage statistics for the forms of N/W+PA across the age groups are shown in
Table 4.27. For four of the ﬁve forms of N+PA, the differences in use among the age
groups are signiﬁcant or approach signiﬁcance. The trend noted previously that the 55+
years group has the least use of N+PP is found for the forms of N+PA, as well. For every
form of N+PA, the 55+ years group ranks the lowest, and these gaps in use among the
age groups are responsible for the signiﬁcance or near signiﬁcance of the differences in
the age distribution of need by, need down, need in, and need out. The differences in the
use of only one form of W+PA approach signiﬁcance across the age groups. For want out
(p=.069), the near Signiﬁcance is perhaps related to its comparatively low use by both the
55+ years group and the 35-54 years group.

Another observation has to do with the difference in use of need by and want by.
Overall, the differences in use of need by (80%) and want by (77%) were not signiﬁcant
in the Midland; however, in the 25-34 years group and in the 35-54 years group, use of
need by (82% and 90%, respectively) is much greater than use of want by (64% and 70%,
respectively), while in the 55+ years group the use of need by (50%) is much less than the

use of want by (100%). Furthermore, there is no discernible preference in the use of the

115

forms of N+PA over W+PA, except in the 55+ years group and in the 35-54 years group

for need oﬂ(73%) and want 0ﬂ(93%).

Table 4.27. Use of N/W+PA by age

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Midland N 18-24 yrs 25-34 yrs 35-54 yrs 55 +yrs Pears“
1? df P
needby 64/11/10/10 83%(53) 82%(9) 90%(9) 50%(5) 6.6 3 .086
need down 64/11/10/10 86%(55) 82%(9) 90%(9) 50%(5) 8.1 3 1.0443-
need in 128/22/20/20 90%(115) 82%(18) 80%(16) 70%(14) 6.7 3 .0847?
needoffl8 97/16/15/16 80%(78) 69%(11) 73%(11) 63%(10)3.2 3 .365
need out 128/22/20/18 82%(105) 77%(17) 75%(15) 56%(10)6.6 3 90847
wantby 64/11/10/10 77%(49) 64% (7) 70%(7) 100%(10) 4.4 3 .225
wantdown 64/11/10/10 92%(59) 91%(10)100%(10)100%(10)1.7 3 .397
wantin 128/22/20/20 85%(109) 77%(17) 90%(18) 90%(18) 1.8 3 .608
wanton“19 97/16/15/15 87%(84) 75%(12) 93%(14) 80%(12) 2.6 3 .459
want out 128/22/20/20 84%(108) 77%(17) 65%(13) 65%(13) 7.1 3 -069”

 

4.4.5. Discussion

The results of the nonlinguistic variables of rural-urban residence, gender, socioeconomic

status, and age were presented in the preceding subsections, and their relevance to the

ﬁeld and suggestions for future research are explored here.

Rural-urban residence

No Si gniﬁcant differences were found in the distribution of the linguistic variables

between rural and urban respondents in the North, the Midland, and the South. The lack

 

 

’8 See footnote 1.
’9 See footnote 2.

116

of rural-urban distinctiveness for NIW+PP and N/W+PA contrasts with the ﬁndings of
DARE for the use of want of in the United States (see Figure 2.1); however, it parallels
Hartman’s (1966:72) ﬁndings for want oﬂ in a southern Ohio county as well as the
ﬁndings of other grammatical variables common in the Midland. No distributional
differences among rural and urban respondents were found in recent studies of
need/ivant/like + past participle, as in 25-27 (Frazer et al. 1996:265-266; Murray &

Simon 19992150, 2002:53).

(25) The house needs painted.
(26) The kids want fed.

(27) The dog likes petted.

Several reasons may account for the differences in the ﬁndings of DARE as
compared to those of this study and Hartman (1966). One reason may have to do with the
treatment of the variables ‘rural’ and ‘urban’. DARE (Cassidy 1985zxiii) deﬁned ﬁve
community types from urban to rural based on population and locality (neamess to a rural
or urban area), while the present study and Hartman’s study, had only two community
types. In the present study, rural was quite narrowly deﬁned (see Section 3.2), equal to
DARE ’3 ‘rural’ and ‘village’ community types, while urban was much more broadly
deﬁned, encompassing DARE is ‘small city’, ‘large city’, and ‘urban’ community types.
Nevertheless, these differences between two and ﬁve community types do not seem to
account for the differing ﬁndings. Combining DARE ’s community types to better match

those in the present study results in use of want of by 22% (N=670) of rural dwellers and

117

12.6% (N=317) of urban dwellers. These percentages of use are not only much less than
any of the linguistic variables in the present study but also the ratio of rural to urban use
is much greater in DARE than in the present study.

Moreover, DARE tended to oversample the lower two community types, N=670,
while undersampling the upper three community, N=317, (see Figure 2.4), while the
present study has a greater number of urban respondents, N=121, than rural respondents,
N=46. These differences in sampling methods may account for the differences in use
among rural and urban respondents. Another reason for the difference may have to do
with the fact that the DARE statistics on want of were calculated for use in the entire
United States, while the statistics in Section 4.4.1 were calculated within each dialect
region (i.e. the North, the Midland, and the South) and in Hartman’s (1966) were based
only on data from one county in southern Ohio, an area where want of has been shown to
be common. A fourth reason may have to do with the passage of timeand the prospect of
language change. The DARE data was collected between 1965-1970, some 30 years
before the present Study. AS other social variables seem to indicate, these linguistic
variables may be undergoing language change and gaining acceptance in rural and urban
areas alike.

Based on Hartman’s ﬁndings on want oﬂ and on studies of needxivant/like + past
participle (as illustrated in 25-27), it may not be so surprising that the forms of NIW+PP
and N/W+PA do not Show any rural-urban distinctiveness in the Midland; however, these
forms also do not Show any rural-urban distinctiveness in the North and the South, which
might be expected if these forms are indeed in the process of Spreading to these areas.

These ﬁndings are important to the ﬁeld of language variation and change because they

118

suggest that grammatical variables may behave differently from lexical and
phonetic/phonological variables with respect to rural-urban distribution, particularly in
cases where the linguistic variable appears to be well-established within one area and is
likely spreading into other areas.

Numerous examples of rural-urban distributional differences of lexical and
phonetic/phonological variables can be found in DARE. Moreover, Labov and others
have deﬁned two major sound changes currently in progress in the United States by their
roles in rural and urban areas. As the name indicates, the Northern Cities Chain shift is
well established in urban areas and only recently has started to reach smaller cities and
has yet to gain a foothold in less urban areas (Eckert 1991; Labov et a1 1997; Ito 1999;
Labov 20012285; Labov et al to appear). On the other hand, the Southern shift is well
established in less urban areas and is not well advanced in more urban areas (Fridland
1998; Labov et al 1997, to appear). The fact that N/W+PP and N/W+PA appear to be
spreading to the North and South without a hint of rural-urban distinctiveness suggests
that these grammatical variables are not following in the footsteps of many a

phonological and lexical variable.

Gender

The gender differences in the use of the forms of N/W+PP and NIW+PA are relatively
unremarkable in the North and the Midland; however in the South, they are intriguing. In
the North and the Midland, only one form Showed Signiﬁcant differences in use by
gender: want oﬂ+ NP was used signiﬁcantly more by men than by women in the North,

while need out was used signiﬁcantly more by men than women in the Midland. In the

119

South, however, four forms Showed Signiﬁcant differences in use by gender—need in +
NP, want in + NP, need in, and want oﬂ—and three others had differences that
approached Signiﬁcance—need oﬂ+ NP, need out, and want in. In every case, use by
men exceeded use by women.

Many of these differences, or lack of differences, in the use of NIW+PP and
N/W+PA by gender may be attributable to an inequity in sampling. Overall, men
outnumber women in the sample 114 to 49; within each region, the breakdown is as
follows: the North (31 to 10), the Midland (64 to 31), and the South (19 to 8). Although
statistical programs are designed to compensate for sampling inequities, they can best
compensate when the sample sizes are large, unlike those in the North and the South.

Nevertheless, it seems that the use of N/W+PP and N/W+PA is more advanced
among men than women in the South. Research on the role of gender in language
variation and change has revealed several tendencies, including the following: Women
tend to use more standard language features, while men tend to use more nonstandard
features; and women tend to more readily adopt new prestige language features than men
(e.g. Labov 1966:312-313; Chambers & Trudgill 1980:72-73; Eisikovits 1989:37-41; see
Labov 2001:261-293). In the South, N/W+PP and N/W+PA must not be viewed as
overtly prestigious incoming forms, since men rather than women lead in the use of these
constructions. Furthermore, NIW+PP and N/W+PA very likely at some level have covert
prestige attached to their use, which might explain the greater use of these constructions
by men. Moreover, these constructions may have no positive or negative evaluation
attached to their use, so women are avoiding them, while men are picking them up.

Certainly, more research in communities in the South is needed to determine if change is

120

indeed taking place, if these forms are evaluated negatively or positively, and if men are

indeed leading women or women are retarding their use of the constructions.

Socioeconomic status

The patterning of the use of the forms of NIW+PP and NIW+PA aCross the
socioeconomic classes shows little variation in the Midland. Only want out showed

Si gniﬁcant differences in use by socioeconomic status, being preferred more by the lower
middle and upper middle classes and much less by the upper working and lower working
classes. The ﬁndings that nearly all forms of NIW+PP and NIW+PA are not stratiﬁed by
socioeconomic class parallel the ﬁndings of other grammatical variables with strong
Midland distributions. For example, needxtvant/like + past participle (illustrated in 25-27)
and positive anymore, illustrated in 28, have been shown to resist socioeconomic
stratiﬁcation in the Midland (e.g. Labov 1991:277; Murray 1993:185; Frazer et al

1996:264-265; Murray & Simon 1999:149-150, 2002:53, 55).
(28) We go to the movies a lot anymore. (53)

It would seem that within the Midland, the use of N/W+PP and N/W+PA is well-
entrenched and not in any way socially stigmatized, as evidenced by the lack of
differences in use across socioeconomic classes. Further study of the attitudes toward the
use of and users of NIW+PP and NIW+PA is needed.

This lack of socioeconomic signiﬁcance seems to parallel the lack of gender

differentiation in the Midland: Since no clear ‘prestige’ form is evident, these forms are

121

not avoided by men or preferred by women. The gender signiﬁcance in the South
suggests that these forms may be treated differently there; however, the lack of
information on the socioeconomic stratiﬁcation in the South does not allow for further
consideration at this point. The relevance of the lack of socioeconomic differentiation is
further discussed in the Section 4.5.2 in relation to stylistic variation.

In Section 4.4.3, it was observed that lower middle class respondents tend to be
among the greatest users of NIW+PP and N/W+PA and that upper middle class
respondents tend to be mid-range users of these variables (i.e. never greatest, never least).
Such patterns are typical of phonological variables undergoing language change. It may
be the case that even within the Midland increased prestige is being attached to these
forms and that the greater use by lower middle respondents is evidence of a change in
progress. As observed in Section 4.1.3, although the use of N/W+PP and N/W+PA is
common in the Midland, it is not a given. Even within the same communities, there are
users and nonusers. Therefore, there is room for N/W+PP and NIW+PA to spread within
the Midland. Additional study is needed to investigate the attitudes toward NIW+PP and

N/W+PA and to explore the idea of a change in progress within the Midland.

Age

Finally, the ﬁndings on the use of NIW+PP and N/W+PA with respect to the fourth social
variable, age, are discussed. The hypothesis that the use of N/W+PP and NIW+PA would
not vary by age was blown out of the water by the behavior of the oldest age group. The
respondents in the 55+ years group tended to use the forms of NIW+PP and N+PA much

less than respondents in the three younger age groups. For need in + NP, want in + NP,

122

need by, need down, need in, and need out, these differences were signiﬁcant or
approached signiﬁcance. The use of the forms of W+PA by the respondents in the 55+
years group is generally much greater and tends to be in line with that of the other age
groups, with the exception of want out. These differences across the age groups, the
differences in the behavior of the 55+ years group, and the general usage statistics, all
suggest that the forms of W+PA, in particular, want by, want down, want in, and want of,
are the most used and oldest forms of NIW+PP and N/W+PA. For reasons unknown at
this point, want 017+ NP also appears to belong to this group. Nevertheless, the other
forms of N/W+PP and N+PA are not far behind, as illustrated by their considerable levels
of acceptance among the other three age groups. Further study of the use of N/W+PP and
N/W+PA across age groups is needed, particularly within one community in the Midland,
to substantiate the conclusionsidrawn here.

This section cannot conclude without some discussion of the behavior of want
out. Frankly, I am puzzled by the differences in use found for want out across
socioeconomic classes and age as well as its relatively lesser rate of use as compared to
other forms of W+PA in the Midland (shown in Table 4.4). Previous literature had
suggested that want out was among the most used forms of W+PA, and yet, the two
lowest socioeconomic classes and the two older age groups do not seem to like this form
as well as other forms. The lesser rate of acceptability of want out seems to lie in the
semantic distinction of concrete and abstract: The differences in the use of abstract senses
of want of by socioeconomic status and by age were Signiﬁcant, owing to lesser rates of
use among the lower socioeconomic classes and the older age groups, while the

differences in the use of concrete senses by socioeconomic status and by age were not

123

signiﬁcant. The only explanation I can suggest for these ﬁndings is that the respondents

in these groups did not respond well to the token sentences for abstract want out in 29-30.

(29) This marriage is over. I want out. (48)

(3 O) This relationship is going nowhere. I want out. (47)

First, the term ‘relationship’ has only more recently become a widely used term for
couplings resulting from dating or marriage; thus, it may not have been as acceptable to
those in the older age groups and in the lower socioeconomic classes. Second, there is
perhaps greater respect for the institution of marriage, especially among older and more
traditional respondents, such that the idea of ‘wanting out’ of a marriage is unacceptable.
In summary, there is a lack of variation in the use of NIW+PP and NIW+PA by
rural-urban residence, gender, and socioeconomic status in the Midland, supporting the
hypotheses. There is, however, some variation by age, especially for NIW+PP and
N+PA, which showed lesser use among the oldest respondents. In the North and the
South, only the social variables rural-urban residence and gender were examined, and
there was no variation in use by rural-urban residence, though there was by gender. In
general, the use of NIW+PP and NIW+PA was greater among men than women, and this
was most significant in the South. On the one hand, the ﬁndings Show little or no social
variation, particularly in the Midland and the North; however, where there is variation, it
seems to hint at a change in progress. Additional research is needed in communities in the
North, the Midland, and the South to investigate the possibility of change and to compare

the characteristics of grammatical change with phonetic/phonological change.

124

4.5. Stylistic variation
In this section, the use of N/W+PP and NIW+PA across stylistic levels is examined.
J oos’s (1967) classic discussion of style identiﬁed ﬁve stations on the formal to informal

style continuum, as shown in (31).

(31) most formal < >most informal
frozen formal consultative casual intimate

 

At the opposite ends of the scale, ‘frozen’ and ‘intimate’ styles are the most limited in use
and range and also the most stable (Joos 1967 :39, 30). While frozen style is ‘for print and
for declamation’ and lacks interaction on the part of the heater, intimate style is particular
to a close-knit group, with much shared background knowledge, and is characterized by
ellipsis, jargon, and a lack of Slang (J 008 1967 :39, 30-32). Of the styles between the most
informal and the most formal, because casual style is used with friends and
acquaintances, its use assumes some shared background knowledge and is characterized
by the use of ellipsis and slang (Joos 1967 :23). Consultative style lies in the middle of the
continuum and is the style used with people one does not know or does not know well.
The use of consultative style assumes no background knowledge; however, it does
assume hearer participation. Finally, formal style, which lies between consultative and
frozen styles on the continuum, has the main purpose of informing. The deﬁning
characteristics of formal style are ‘advance planning’, ‘detachment’, and ‘cohesion’ (Joos
1967 :35, 37-3 8). On the questionnaire used in this study, there were four categories of

usage responses, ‘b’, ‘c’, ‘d’, and ‘e’ (as discussed in Section 3.2):

125

Choose b if you would use this sentence only in writing or in very formal Speech
Situations like a job interview or in court.

Choose c if you would use this sentence only with close friends and/or family.
Choose d if you would use this sentence in general conversation, in stores and with
people you don’t know well.

Choose e if you would use this sentence in all situations.
The questionnaire responses relate to J oos’s continuum as follows:

‘b’, labeled ‘formal only’, covers formal and some aspects of frozen styles;
‘d’, labeled ‘consultative only’ covers consultative styles;
‘c’, labeled ‘informal only’, covers casual and some aspects of intimate styles;

‘e’, labeled ‘all use’, covers all styles (‘e’).

In the following subsections, the results of the use of NIW+PP and NIW+PA

across stylistic levels are presented.

4.5.1. NIW+PP

Table 4.28 presents the use of every token of NIW+PP by style. Overall use of N+PP and
W+PP are similar across stylistic levels. For N+PP and W+PP, use in all styles is greatest
(34% and 39%, respectively), followed by use in informal style only (24% and 27%,
respectively), consultative style only (25% and 24%, respectively), and formal style only

(17% and 4%, respectively).

126

Table 4.28. Use of all tokens of N/W+PP by style

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Midland Use Formal Consultative Informal All

11 Only Only Only Use
Overall use of N+PP 291 17% (50) 25% (73) 24% (69) 34% (99)
(1) Dean needs in the building. 40 8% (3) 28% (11) 28% (11) 38% (15)
(2) This box needs in the car. 25 4% (1) 24% (6) 56% (14) 16% (4)
f:)tf:;:i‘::ewj°b1 Tony "9““ 39 23% (9) 15% (6) 23% (9) 38% (15)
(4) Katy needs in this class in o o o 0
order to graduate this semester. 41 15 /° (6) 27A) (11) 15 /° (6) 44/0 (18)
(5) Peter needs of] the plane. 37 11% (4) 32% (12) 19% (7) 38% (14)
E20?” b°xes need 017 the 31 0% (0) 45% (14) 29% (9) 26% (8)
27f} $312332?” and ”93‘“ 37 14% (5) 22% (8) 24% (9) 41% (15)
(8) The lawyer had a conﬂict of o o o 0
interest and needed of the case. 41 54 A) (22) 12 /° (5) 10 A (4) 24 /° (10)
Overall use of W+PP 339 4% (15) 24% (82) 32% (110) 39%(132)
i131): 3:131:23; :1” the $an use 46 4% (2) 28% (13) 35% (16) 33% (15)
(10) When the cat wants in the o o o 0
house, she scratches at the door. 40 3 A) (I) 13 /° (5) 50A (20) 35 /° (14)
(IigsThis student wants in your 42 5% (2) 21% (9) 24% (10) 50% (21)
(12) Sam wanted in the Army. 34 9% (3) 18% (6) 47% (16) 26% (9)
(13) I want oﬂthis horse. 44 0% (0) 18% (8) 39% (17) 43% (19)
2:301:23?“ wa’" Off’he mm" 44 2% (1) 32% (14) 27% (12) 39% (17)
(15) Monica lost interest in
playing baseball and wanted of 46 7% (3) 15% (7) 35%(16) 43% (20)
the team.
(16) I want oﬂ your mailing list. 43 7% (3) 47% (20) 7% (3) 40% (17)

 

 

 

 

 

 

127

 

Two generalizations may be drawn from the data: 1) Forms of NIW+PP are used
in all styles, as evidenced by the generally large number of respondents who indicated
that they would use these sentences in all styles. 2) Forms of N/W+PP are used less in
formal styles than in consultative and informal styles. There are some exceptions to this
generalization, and these exceptions Skew the data in such a way that N+PP appears to be
considered more formal than W+PP. In a few cases, the respondents may have been
unduly swayed by the instructions on the questionnaire. For example, 54% of respondents
who report using sentence 8 would use it in the formal style. Respondents were perhaps
inﬂuenced by the instructions given on the questionnaire, in which ‘court’ was listed as
an example of a formal situation (see Section 3.2 or Appendix B). It seems likely that
some respondents focused on the words ‘lawyer’, ‘conﬂict of interest’, and ‘case’, all
terms that are used in court. In spite of the fact that some sentences may have contained
references that may have led some respondents to determine their stylistic level on the
basis of facts other than those targeted in this study, the hypothesis that N/W+PP is used
more in more informal styles than more formal styles, proposed in Section 2.6, is largely
supported by the data.

Although NIW+PP is used less in more formal situations, it cannot be
characterized as strictly ‘informal’ for two reasons. As noted previously, a substantial
percentage of respondents report that they would use N/W+PP in all situations.
Furthermore, there is quite a bit of variation between exclusive use in informal styles and
consultative styles, a less informal. (or more formal) style. For some sentences, there is
greater reported use in informal styles, e. g. sentences 2, 10, and 12, while for others, there

is greater reported use in consultative style, e. g. sentences 6 and 16.

128

4.5.2. N+PA
Table 4.29 presents the use of every token of N+PA by style. The forms of N+PA behave
much like those of N/W+PP. Overall, use in all Styles is greatest (34%), followed by use
in informal style only (3 2%), consultative style only (26%), and formal style only (9%).
Again, two generalizations emerge here: 1) Forms of N+PA are used in all styles,
and 2) they are used less in formal styles than in consultative and informal styles. In
general, the hypothesis, proposed in Section 2.6, that N+PA is used more in more
informal styles rather than more formal styles is supported by the data. Here, too,
although N+PA is used less in more formal styles, it is not strictly informal. A number of
respondents indicate that they would use these sentences in all styles, and there is much
variation between reported use in informal styles and consultative styles. For some
sentences, there is greater reported use in informal styles, e. g. 22, 24, and 30, while for

other sentences, there is greater reported use in consultative styles, e. g. 25 and 26.

129

Table 4.29. Use of all tokens of N+PA by style

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

out.

 

 

 

 

 

Use Formal Consultative Informal All Use
Midland 11 Only Only Only
tOverall use of N+PA 605 9% (52) 26% (1.59) 32%(191) 34%(203)
£1322? 5°“? ‘° be‘he‘ ye“, but I 39 8% (3) 38% (15) 18% (7) 36% (14)
2;” Step heel" Please Anne needs 37 5% (2) 30% (11) 19% (7) 46% (17)
(19) The baby is on the couch and o o o 0
needs down. 41 2A; (1) 24/o (10) 37/6 (15) 37/o (15)
(20) The cat is stuck in the tree and o o o 0
needs down. 37 3 /o (1) 38 /o (14) 35 /o (13) 24/6 (9)
(21) The cat needs in. 43 9% (4) 14% (6) 37% (16) 40% (17)
(22) The door is locked and I need in. 42 2% (1) 12% (5) 48% (20) 38% (16)
(23) The day care center is full, and o o o o
my child needs in. 41 7/o (3) 29A; (12) 24/o (10) 39/6 (16)
(24) The new drug experiment is o o o o
Monica’s last hope. She needs in. 37 19/0 (7) 5 /° (2) 46 /° (17) 30 A) (11)
(25) The fumes from this bus are 0 o o 0
making me sick. I need off. 44 7A1 (3) 43 /o (19) 23 /o (10) 27A) (12)
$3123: 3;“ is e“ the wreng Plane 31 6% (2) 39% (12) 13% (4) 42% (13)
(27) I’ve wasted too much time work- 0 o o 0
ing on this project, and I need off. 31 13 A) (4) ’ 26/0 (8) 32A) (10) 29/0 (9)
(28) There are too many problems
with the dance committee, and Jill 35 11% (5) 31% (11) 37% (13) 17% (6)
needs off
(122:) The dog needs out three tlmes a 40 5% (2) 17% (7) 32% (13) 44% (18)
(30) Mike needs out at the corner. 35 3% (1) 29% (10) 43% (15) 26% (9)
(31) Barry’s job has become too 0 o o o
stressful and he needs out. 35 14/o (5) 23 /o (8) 34/6 (12) 29/o (10)
(32) The company agreed to a deal ~
two weeks ago, but now they need 37 22% (8) 24% (9) 24% (9) 30% (11)

 

130

 

4.5.3. W+PA

The use of all tokens of W+PA by style are presented in Table 4.30. Overall, use in
informal style only is greatest (36%), followed by use in all Styles (34%), consultative
style only (22%), and formal style only (7%). This overall generalization favoring the use
of the forms of W+PA in informal styles is perhaps overstated by a few sentences with
disproportionately greater uses in informal styles, e. g. 44, 46, and 48. The hypothesis that
W+PA is used more in more informal styles than more formal styles, proposed in Section
2.6, is supported, though not overwhelmingly.

AS noted previously, some of the disproportionately greater uses in informal style
were likely due to some respondents’ sensitivity to content as well as their adherence to
the descriptions for each answer choice. For example, in sentences 47 and 48,
respondents recognize based on the content that these sentences would likely only be
uttered to those identiﬁed in answer choice ‘c’, i.e. close friends and family.
Nevertheless, the generalizations found in the data for the forms of NIW+PP and N+PA
also apply to the forms of W+PA. The forms of W+PA are indeed used in all styles, and
the use in more formal styles is much less than in other styles.

Overall, the use of N/W+PP and N/W+PA leans toward the informal side of the
formality continuum, supporting the hypothesis, proposed in Section 2.6. This does not
mean, however, that N/W+PP and N/W+PA are used exclusively in informal styles. For
nearly every token sentence, a substantial number of respondents would reportedly use
the sentence is all styles. How formal or informal these constructions are is not entirely
clear, since there is considerable variation in the reported use of individual token

sentences at the consultative only and informal only styles.

131

Table 4.30. Use of all tokens of W+PA by style

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

out.

 

 

 

 

 

. Use Formal Consultative Informal
Midland n Only Only Only All Use
Overall use of W+PA 614 7% (43) 22% (137) 36%(224) 34%(210)
(33) Mark wants by, but he is too 0 50 o o
Shy to ask people to move. 37 8/0 (3) 3 /° (13) 16/0 (6) 41/0 (15)
(34) I can move ifyou want by. 36 6% (2) , 31% (11) 25% (9) 39% (14)
(3 5) Do you want down? 46 7% (3) 20% (9) 22% (10) 52% (24)
(36) The baby wants down. 43 2% (1) 21% (9) 44% (19) 33% (14)
(37) The kids say it’s too hot 0 50 36° 6 4o 15
outside, and they want in. 44 5 /° (2) 2 /° (1 1) /° (1 ) 3 /°( )
(3 8) When the dog scratches at the o o o 0
back door, he wants in. 42 5 /o (2) 21 /o (9) 48 A: (20) 26 At (11)
(3 9) This project is exactly what 0 ' o o o 4
Donna likes to do. She wants in. 37 14A (5) 22A) (8) 27A (10) 38 /° (1 )
ngtfgat sounds like a great plan. I 39 5% (2) 15% (6) 3 6% (14) 44% (17)
(41) When the roller coaster picked
up Speed, Ed yelled that he wanted 45 11% (5) 31% (14) 27% (12) 31% (14)
01f-
(42) The child was scared on the o o o 3 o 2
swing and said she wanted 017. 39 10/0 (4) . 26/0 (10) 33 /° (1 ) 31A, (1 )
(43) The baseball team hasn’t won a
game in several weeks, and Jim 38 11% (4) 29% (11) 39% (15) 21% (8)
wants oﬂf
(44) Stop the world! I want oﬁf 17 6% (1) 6% (l) 53% (9) 35% (6)
(45) Dan was getting nervous in the o o o o
stuck elevator. He wanted out. 40 10A, (4) 25 A (10) 25 /° (10) 40 /° (16)
$212123?“ want out whenever it’s 38 0% (O) 13% (5) 61% (23) 26% (10)
(47) This ’eleﬁenSh’p is 80mg 38 5% (2) 16% (6) 45% (17) 34% (13)
nowhere. I want out.
(48) This marriage is over. Iwant 35 9% (3) 11% (4) 60% (21) 20% (7)

 

132

 

4.5.4. Discussion

In this section, the signiﬁcance of the stylistic variation of N/W+PP and N/W+PA,
reported on in the previous sections, is discussed. In general, two tendencies are found
with respect to the use of N/W+PP and NIW+PA by style: 1) These constructions are
reportedly used in all styles, and 2) they are reportedly used more in more informal styles
than in more formal styles, supporting the hypothesis proposed in Section 2.6. The ﬁrst
generalization is supported most basically by the fact that of the number of respondents
who report using the token sentences at all, more than one-third would use them in all
styles. These statistics taken together with the number of respondents who would use
NIW+PP and N/W+PA only in formal and consultative styles refute any notion in the
previous literature that these constructions are appropriate only to informal styles (e. g.
Bryant 19622224; Quirk et a1 1985:732; Greenbaum & Whitcut 19882762; K. Wilson
1993). This generalization is important for an accurate description of the use of N/W+PP
and N/W+PA.

Perhaps even more interesting is the variation in the use of N/W+PP and
N/W+PA by style, as illustrated in Figure 4.5. The reported use of these constructions
increases in proportion to the informality of the situation. Although precisely this type of
stylistic variation was predicted in this study, it is not in-line with current sociolinguistic
theory. The widely accepted belief, particularly among variationists, with respect to
stylistic variation is that there is a direct relationship between the social distribution of a
linguistic variable and its stylistic variation (e.g. Romaine 19802228; Labov 1964:101-
102, 1966:222-223, 1972:126; Wolfram & Fasold 1974:85-86; Bell 19842146; Finegan &

Biber 1994:315-316).

133

40% ~
35% ‘

30% ﬂ /

25% i I

 

20%1
15% ~
10% ~

5%‘

 

 

0% 1 , ,
Formal only Consultative only Informal only

[+N+PP +W+PP +N+PA -D—W+PA

 

 

Figure 4.5. Use of NIW+PP and NIW+PA by style

Moreover, as a derivative to his style axiom (discussed in Section 2.6), Bell
(1984:152) insists that for a given variable, its socioeconomic variation MUST be greater
than its stylistic variation: ‘Where style shift does occur, it is notable that, quantitatively,
the degree of style variation never exceeds the degree of social variation.’ Preston
(1991 :35) goes even ﬁrrther by ‘deﬁning the variation space of the stylistic dimension not
only as smaller than but also as contained within that of the social dimension.’ Bell and
Preston are not alone in their assertions. The much used categories of sociolinguistic
variables—stereotypes, marker, and indicators—exclude the case of variables with
stylistic but little or no socioeconomic variation (see e. g. Labov 1972: 178-179; Wolfram

& Fasold 1974:83-87; Downes 1998:188-196; Wolfram & Schilling-Estes 1998:161-

134

162). In fact, current theory holds that stylistic variation not only cannot exceed social
variation but MUST be smaller than social variation.

In this study, such claims are turned on their heads. As seen in Section 4.4, in the
Midland, N/W+PP and N/W+PA do not show any variation not only across
socioeconomic classes but also other social variables—rural-urban residence, gender, and
age—yet these constructions do show variation across styles.

This study is not the ﬁrst to ﬁnd a pattern of stylistic distribution independent of
the social distribution; indeed, Bell (1984:155-156) cites the case of lael/ raising in
Tehrani Persian, which showed enormous stylistic variation between ‘reading’ style and
‘careful speech’ style but little social variation. A social climate requiring ‘extreme
deference’ is apparently what underlies such ‘extreme audience directed style shiﬁ’,
which Bell postulates is also likely to be found in other non-Western societies (Bell
1984: 156). The case here is neither the result of ‘extreme deference’ nor is it situated in a
non-Western context.

Why is this apparently one of the ﬁrst studies to show stylistic variation
independent of social variation that cannot be attributed to some other unusual or extreme
factors? There are several reasons for this. 1-) Much of the data underlying sociolinguistic
theory and its tenets are of phonetic/phonological phenomena. The ﬁeld of
sociolinguistics has not rejected the study of grammatical variation by any means, but a
look in any introductory textbook or academic journal of language variation will reveal
that the number of studies examining grammatical variation pales in comparison to the
number of studies that focuses on phonetic/phonological variation (e. g. Chambers &

Trudgill 1991:291). 2) The paucity of work of grammatical variables is further

135

compounded by the lack of work on stylistic variation and, consequently, stylistic
variation of grammatical variables (e.g. Bell 19842146; Finegan & Biber 19942319). 3)
Perhaps the most important reason why so few studies have appeared that show stylistic
variation without social variation is the power of the long-held assumption encapsulated
in Bell’s style axiom. Most researchers probably decided that if little or no social
variation existed for a linguistic variable, they should not even bother to look for stylistic
variation. If more studies of stylistic variation and grammatical variables are undertaken,
especially among those variables that have shown little or no social variation, e. g.
needﬁvant/Iike + past participle and positive anymore (illustrated in 25-28), perhaps
additional data will be found to support the existence of linguistic variables with stylistic
variation independent of social variation.

Three patterns of the interaction of social and stylistic variables were discussed

and illustrated in Section 2.6:

(l) Sociolinguistic markers—characterized by both social and stylistic stratiﬁcation,
illustrated in Figure 2.5.

(2) Sociolinguistic indicators—characterized by social stratiﬁcation with little
stylistic variation, illustrated in Figure 2.6.

(3) Hypercorrection—characterized by one social class’s use of a variant that crosses

over that of another class’s along the style dimension, illustrated in Figure 2.7.

A fourth pattern must now be added.

136

(4) Stylistic indicators—characterized by stylistic stratiﬁcation with little social

variation, similar to the social and stylistic distribution of W+PP, illustrated in Figure 4.6.

 

 

 

 

 

 

 

 

50% 1

40% r

30% ~ - - - Lower working
— - - Upper working

Lower middle

-— Middle middle

20% i - -Upper middle

10% A

0% r 1
Formal only Consultative only Informal only

Figure 4.6. W+PP by socioeconomic class and style (x2=24, (if 16, p=. 100)

In summary, although the forms of N/W+PP and NIW+PA show little variation
across the social variables of socioeconomic status, rural-urban residence, gender, and
age, discussed in Section 4.4.5, they do show variation across style. NIW+PP and
N/W+PA are used less in more formal styles than in less formal styles, supporting the
hypothesis. The implications for sociolinguistic theory are striking, for it has long been
assumed that stylistic variation cannot be greater than social variation. It is likely that
stylistic variation may exceed social variation in more cases than previously thought if

additional research is canied out, particularly on grammatical variables that have shown

137

little or no social variation. These ﬁndings call for a reevaluation of this long held

assumption.

138

5. Conclusion

This dissertation has examined the use of N/W+PP and NIW+PA not only to determine
their geographic, social, and stylistic distributions but also toexplore other factors that
may affect the distributions of the individual forms of the constructions.

The aims of this ﬁnal chapter are threefold. First, in Section 5.1, I list the general
ﬁndings of this study. Second, in Section 5.2, I present what I consider to be the most
important limitations of this study, and I suggest avenues for future research. Finally, in
Section 5.3, I discuss the more general relevance of this study’s ﬁndings to the ﬁeld of

linguistics.

5.1. Summary of findings
Several interesting ﬁndings about the geographic, social, and stylistic distribution of the

grammatical constructions N/W+PP and NIW+PA have been uncovered in this study.

(1) N/W+PP and N/W+PA have robust distributions in the Midland (see Section 4.1).
(2) In the North and the South, N/W+PP and N/W+PA are not unknown. Previous
research suggested that these constructions were little used in those areas (see Sections
2.2 and 2.3); however, this study found that some forms, e. g. want down, want in, want
01% want out, and want 0ﬂ+ NP were used by over 60% of both Northern and Southern
respondents (see Section 4.1).

(3) The use of N/W+PP and N+PA, in general, was much greater than anticipated,

especially in the Midland, but also in the North and the South (see Section 4.1), since

139

little mention of these constructions had been made in previous literature (see Section
4.1).

(4) There was a clear preference for want over need in PP constructions in all three
dialect areas and in PA constructions in the North and the South (see Section 4.3).

(5) The semantic features of concrete and abstract, as applied to the interpretation of
the “goal” of the construction, i.e. the location identiﬁed by the prepositional
phrase/preposition, appear to play a role in the distribution of NIW+PP and N/W+PA (see
Section 4.2); however, this role is not yet well understood.

(6) NIW+PP and NIW+PA were not used differently by rural and urban dwellers in
any of the three dialect areas (see Section 4.4.1), as might have been predicted on the
basis of previous variationist studies of phonetic/phonological variables.

(7) In the Midland, there were few differences in the use of NIW+PP and N/W+PA
by gender. In contrast, in the South and to some extent the North, these constructions
were used more by men than women (see Section 4.4.2).

(8) Overall, there were few differences in the use of NIW+PP and NIW+PA across
socioeconomic classes in the Midland (see Section 4.4.3).

(9) In the Midland, there was generally little variation in the use of NIW+PP and
N/W+PA, particularly among three of the four age groups. The oldest age group,
however, tended to use the forms of N+PP and N+PA much less than the others (see
Section 4.4.4).

(10) An analysis of the use of NIW+PP and NIW+PA by style revealed that although
these forms are not restricted to use in the informal register, as previous literature had

suggested (see Section 2.6), they, nevertheless, vary by style. In fact, there is gradient

140

stratiﬁcation by style, with less use in the formal register, moderate to greater use in the

consultative register, and greater use in the informal register (see Section 4.5).

The signiﬁcance of these ﬁndings is disCussed in Section 5.3. First in Section 5.2,

some limitations of the present and suggestions for future research are explored.

5.2. Limitations and suggestions for future research

In this section, some limitations of the present study and avenues for future research are
presented. The data in this study was based solely on self-reported responses to
questionnaires, though a limited number of respondents were interviewed to gather
additional demographic and ethnographic data. As discussed in Section 3.3, there is
some debate about which methods of data collection, e. g. self-reports or observed uses,
offer the most valid or reliable measures of actual linguistic behavior. As in many
sociolinguistic studies that investigate elements above the phonological and
morphological levels, the collection of adequate authentic conversational data for
statistical analysis of any sort is nearly impossible.

Another limitation of this study is the lack of ethnic, age, educational, and cultural
diversity of the respondents. This study relies heavily on data from university and college
students, most of whom were young (18-24 years old) and European-American. In the
Midland, 15 additional respondents over the age of 30 were recruited in an attempt to
provide some balance in the respondent pool. Further, the gender bias in the respondent

pool (7 0% men to 30% women, overall) places some limitations on the interpretation of

141

the gender ﬁndings, in particular, those discussed in Section 4.4.5. In light of the
limitations of the present study, several avenues of future research are evident.

First, a large-scale dialectological study of a variety of forms of NIW+PP and
N/W+PA is warranted to assemble a complete picture of the geographic distribution of
these constructions in the United States. Such a study will allow for comparisons of the
behavior of NIW+PP and N/W+PA with other ‘Midland’ grammatical constructions, e. g.
need/want/like + past participle and positive anymore, and will allow for a better
understanding of the composition, inﬂuence, and etiology of the much- contested
Midland dialect area, as a transition zone or a separate dialect area. The difﬁculties in
collecting data on grammatical forms, particularly those with lexical differentiation and
semantic sensitivities, makes this avenue of research a long-term goal.

Second, future research should examine individual communities in the North, the
Midland, and the South. Quantitative and qualitative studies of the use of NIW+PP and
N/W+PA and attitudes toward the use of NIW+PP and N/W+PA in communities in these
areas are needed to explore several questions relating to language change and the role of
perception. Are NIW+PP and NIW+PA undergoing change? More speciﬁcally, are these
constructions in the process of spreading from the Midland to the North and the South? If
they are, do the characteristics of the spread of grammatical variables mimic those of
phonetic/phonological variables? Are these constructions positively or negatively
evaluated by speakers? Are these evaluations reﬂected in their distribution across dialect
areas, social variables, and/or stylistic levels? Discussions with some Midlanders
revealed that they were not aware that NIW+PP and NIW+PA were not used by everyone

in all dialect areas, but this study also found that within the same communities there were

142

users and nonusers of these constructions. What differentiates users and nonusers of these
forms within the same community?

Third, there is an area of promising research more theoretical in nature that moves
beyond the conStructions of NIW+PP and NIW+PA. In direct contrast to long-held
assumptions about the relationship between social and stylistic variation, and in
particular, Bell’s style axiom (1984), are the ﬁndings of this study that the degree of
stylistic variation exceeded the degree of social variation. Further research is needed to
discover whether other variables, especially other grammatical variables like
need/Want/like + past participle and positive anymore, show a similar patterning of robust

stylistic variation with little or no social variation.

5.3. Relevance to ﬁeld of linguistics

This section discusses the relevance of the ﬁndings of this study to several areas of
linguistic scholarship, speciﬁcally dialectology, including the status of dialect areas; the
role of semantic factors in the distribution of grammatical variables; language variation
and change, including mental models; and sociolinguistic theory and the relationship
between social and stylistic variation.

In general, this study adds to the body of knowledge of American dialects by
describing the general geographic distribution of various forms of N/W+PP and N/W+PA
constructions. This study also demonstrates that it can be misleading to base the
distribution of a variety of forms, e. g. NIW+PP and NIW+PA, on one exemplar, e.g. want
oﬂ, since some are used more than others (see Section 4.1) Moreover, this study

contributes additional evidence for the existence of a Midland dialect area as a separate

143

and distinct dialect area rather than simply a transition zone between the North and the
South. In addition to historical evidence, this study provides further evidence that, at least
in the United States, N/W+PP and N/W+PA originate in the Midland and are spreading
from the Midland to the North and South, not the other way around, as the transition zone
perspective suggests. 1

Although the investigation of the semantic features of concrete and abstract
senses of the ‘goal’ or locations identiﬁed by the prepositions in the constructions
N/W+PP and N/W+PA could not be conclusively tied to a clean pattern of regional
versus widespread distribution, it revealed differences in the use of such concrete and
abstract senses of the constructions. Another semantic feature that may affect the
geographic distribution of NIW+PP and NIW+PA has to do with the animacy/agency of
the subject of need/want. Although only two such tokens were included in this study,
sentences with inanimate/patient subjects of need/want appear to be much less acceptable
overall and especially in the North and the South, than those with animate/agent subjects,

as shown in Table 5.1.

Table 5.1. Effect of animacy/agency of the subject of need/want on distribution

 

North N Midland N South N

 

 

 

This box needs in the car. 14% (3) 21 54% (25) 46 10% (1) 10
need in + NP (overall) 37% (30) 82 77% (145) 189 35% (19) 54
:32: boxes need °ff the 33% (7) 21 67% (31) 46 10% (1) 10

 

 

 

 

 

 

 

 

need off+NP(overall) 45%(37) 82 78%(149) 19o 39%(21) 54

 

144

 

These ﬁndings demonstrate that semantic factors should be considered in studies
of grammatical variables, since they may affect respondents’ judgments of the
grammaticality/use of the linguistic variables under investigation.

The number of studies of grammatical variation and change pales in comparison
to those of phonological (and even lexical) variation and change. Any study of a
grammatical variable contributes to the body of knowledge of language variation and
change. This study, in particular, has ﬁndings that appear to run counter to accepted
trends in this area of study. For example, it is often noted that the rural-urban distinction
plays an important role in sound change; that is, speakers in rural and urban areas are
often differently engaged in the change, e. g. the Northern Cities Chain Shift, which is
further advanced in urban than in rural areas, and the Southern Shift, which is more
predominant in rural areas than urban areas. In this study, N/W+PP and NIW+PA do not
show any differences in distribution in rural and urban areas, even in dialect areas to
which these constructions appear to be spreading. Furthermore, studies of phonological
variables, and in particular those undergoing change, have shown stratiﬁcation by such
social variables as age, gender, and socioeconomic class. In this study, little to no such
variation is found, particularly in the Midland. These frndings, of course, call for
additional study to substantiate these trends; nevertheless, they raise questions about the
differences between the characteristics of phonological as opposed to grammatical
variation and change.

The usage hierarchies of the forms of NIW+PP and N/W+PA uncovered in this
study present issues for the conceptualization of language variation and change in the

mind. Mental models of language must not only be able to account for variation but also

145

for the fact that some forms, e. g. want down, are far more acceptable than other
seemingly identical forms, e. g. need down or want by (see Table 4.4). Moreover, mental
models must also be able to account for the fact that some forms may be required for the
use of other forms, i.e. have an implicational relationship. For example, it may be the
case the use of need down is contingent on the use of want down at least for some
speakers.

Finally, and perhaps most importantly, this study calls into question long-held
assumptions about the relationship between social and stylistic variation within
sociolinguistic theory. More speciﬁcally, the present study provides counterevidence for
two foundational ideas: 1) Stylistic variation and social variation are linked, e. g. variants
common in more informal styles are characteristic of lower socioeconomic classes, and
2) the range of stylistic variation is never greater than the range of social variation, i.e.
Bell’s style axiom. This study demonstrates-that widely used grammatical variables,
NIW+PP and NIW+PA, with little social variation to speak of are, nevertheless, stratiﬁed
by style.

The ﬁndings of this project have implications for several areas of linguistic study
and it is hoped that additional research will be carried out to further enlighten these

issues.

146

APPENDICES

147

APPENDIX A

Questionnaire items organized by type

148

Table A. 1. Token sentences of N/W+PP

 

Version of Questionnaire

 

Al
Item #

A2
Item #

B1
Item #

B2
Item #

 

Need In + NP

 

(1) Dean needs in the building.

17

48

 

(2) This box needs in the car.

42

15

 

(3) For his new job, Tony needs in the union. (+abstract)

25

21

 

(4) Katy needs in this class in order to graduate this
semester. (+abstract)

34

38

 

Need Off + NP

 

(5) Peter needs off the plane.

56

 

(6) Those boxes need off the truck.

50

 

(7) Randy is too busy and needs off the committee.
(+abstract)

54

ll

 

(8) The lawyer had a conﬂict of interest and needed off the
case. (+abstract)

17

31

 

Want In + NP

 

(9) Ifyou want in the safe, use this combination.

13

54

 

(10) When the cat wants in the house, she scratches at the
door

48

58

 

(1 1) This student wants in your class. (+abstract)

50

 

(12) Sam wanted in the Anny. (+abstract)

31

29

 

Want Off + NP

 

. (13) I want off this horse.

58

19

 

(14) Do you want off the merry-go-round?

50

11

 

(15) Monica lost interest in playing baseball and wanted off
the team. (+abstract)

44

 

 

(16) I want off your mailing list. (+abstract)

 

 

 

36

 

23

 

149

 

Table A2. Token sentences of N+PA

 

Version of Questionnaire

 

Al

' Item #

A2
Item #

Bl
Item #

B2
Item #

 

Need by

 

(17) I’m sorry to bother you, but I need by.

33

28

 

(18) Step back, please. Anne needs by.

46

56

 

Need Down

 

(19) The baby is on the couch and needs down.

60

 

(20) The cat is stuck in the tree and needs down.

46

 

Need In

 

(21) The cat needs in.

34

 

(22) The door is locked, and I need in.

23

34

 

(23) The day care center is full, and my child needs in.
(+abstract)

36

15

 

(24) The new drug experiment is Monica’s last hope. She
needs in. (+abstract)

58

21

 

Need Oﬁ‘

 

(25) The fumes from this bus are making me sick. I need off.

27

4O

 

(26) This man is on the wrong plane. He needs off.

29

 

(27) I’ve wasted too much time working on this project, and
I need off. (+abstract)

48

25

 

(28) There are too many problems with the dance
committee, and Jill needs off. (+abstract)

11

60

 

Need Out

 

(29) The dog needs out three times a day.

21

42

 

(3 0) Mike needs out at the corner.

54

 

(31) Barry’s job has become too stressful and he needs out.
(+abstract)

42

17

 

(3 2) The company agreed to a deal two weeks ago, but now
they need out. (+abstract)

 

 

 

 

13

 

52

 

150

 

Table A3. Token sentences of W+PA

 

Version of Questionnaire

 

Al

5 Item #

A2
Item #

Bl
Item #

B2
Item #

 

Want By

 

(33) Mark wants by, but he is too shy to ask people to move.

3

52

 

(34) I can move if you want by.

15

36

 

Want Down

 

(3 5) Do you want down?

31

46

 

(3 6) The baby wants down.

52

13

 

Want In

 

(3 7) The kids say it’s too hot outside, and they want in.

11

58

 

(3 8) When the dog scratches at the back door, he wants in.

44

 

(3 9) This project is exactly what Donna likes to do. She
wants in. (+abstract)

44

26

 

(40) That sounds like a great plan. I want in. (+abstract)

27

48

 

Want Off

 

(41) When the roller coaster picked up speed, Ed yelled that
he wanted off.

23

32

 

(42) The child was scared on the swing and said she wanted
off.

44

40

 

(43) The baseball team hasn’t won a game in several weeks,
and Jim wants off. (+abstract)

40

 

(44) Stop the world! I want off. (+abstract)

21

54

 

Want Out

 

(45) Dan was getting nervous in the stuck elevator. He
wanted out.

46

23

 

(46) The kids want out whenever it’s not raining.

38

33

 

(47) This relationship is going nowhere. I want out.
(+abstract)

19

6O

 

 

1(48) This marriage is over. I want out. (+abstract)

 

 

 

33

 

17

 

151

 

Table A.4. Filler sentences

 

Version of Questionnaire

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Al A2 B1 B2
Item # Item # Item # Item #

Filler Sentences

(49) I bought myself a new TV. 1 1 1 1
(50) He ugly. 6 6 6 6
(51) I bought ﬁve pound of potatoes. 8 8 8 8
(52) I ain’t going to do it. 10 10 10 10
(53) We go to the movies a lot anymore. 12 12 12 12
(54) He don’t know nothing. 14 14 14 14
(55) The sugar is all. 16 16 16 16
(56) We had went to California on vacation. 18 18 18 18
(57) He throwed the ball at the window. 20 20 20 20
(58) Bill and me went to the game. 22 22 22 22
(59) He don’t know where she lives. 24 24 24 24
(60) He caught himself a cold. 26 33 26 26
(61) Ten o’clock is all the laterI can stay up. 30 31 3O 3O
(62) Look at that dog come a-running. 32 27 32 32
(63) Anymore, we eat ﬁsh often. 35 35 35 35
(64) We be tired. 37 37 37 37
(65) Margaret seen him do it. 39 39 39 39
(66) Janet’s daughter is right smart. 43 43 43 43
(67) Jonathan bought dinner for Mike and I. 47 47 47 47
(68) Bob’s a-ﬁxing the tractor. 49 49 49 49
(69) Do you want to come with? 50 51 51 51
(70) He be singing all the time. 53 53 53 53
(71) They was singing while Jack was working. 57 57 57 57

 

 

 

 

 

152

 

Table A5. Widely accepted ﬁller sentences

 

Version of Questionnaire

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

, Al A2 Bl BZ
Item # Item # Item # Item #

Filler Sentences Accepted Widely
(72) The kitchen smells good. 2 62 2 62
(73) Eddie is very athletic. 4 4 4 4
(74) Where are you going on vacation? 28 29 28 28
(75) I’m not going to do it. 41 41 41 41
(76) They are friendly. 45 45 45 45
(77) I want you to ﬁx the chair right now. 55 55 55 55
(78) I’m not going to paint the house this year. 59 59 59 59
(79) Alice doesn’t eat meat anymore. 61 61 61 61
(80) Mary took Bridget and me to school. 62 2 62 2
(81) We need to get oﬁ" at the next stop. 15 50
(82) Sam wants to get in the Army. 29 3O
(83) Mike says he wants to go off his medication. 34 36
(84) When the cat meows at the door, she wants to go out. 38 5
(85) In stormy weather, the plants on the porch need to come
in. 52 38
(86) Can you help the baby? She wants to get down. 56 13
(87) If you want to get in the apartment while I’m gone,
here’s the key. ~ 40 25
(88) The aisle is crowded, and I need to get by. 60 42
(89) Jenny needs to get in the house. . 9 7
(90) Chris is too busy and needs to get off the team. 25 19
(91) Larry is tired of being in the house and needs to get out. 19 27
(92) I want to get off at Main Street. 56 3

 

153

 

APPENDIX B

Sample questionnaire, version A1

154

Questionnaire A1

Instructions

The following questionnaire is a part of a research project on language use. Your
participation is entirely voluntary, and you indicate your voluntary agreement to
participate by completing and returning this questionnaire.

Please put all responses on the op—scan sheet provided. Do NOT put your name on
the questionnaire or the op-scan form. When completing the questionnaire, please
give your own personal response to the items. There are no right or wrong answers.

Read each of the following sentences to yourself. Rate each sentence on the following
scale:
Choose a if you would never use this sentence.
Choose b if you would use this sentence only in writing or in very formal speech
situations like a job interview or in court.
Choose c if you would use this sentence only with close friends and/or family.
Choose d if you would use this sentence in general conversation, in stores, and with
people you don’t know well.
Choose e if you would use this sentence in all situations.

1. I bought myself a new TV.

2. The kitchen smells good.

3. Mark wants by, but he’s too shy to ask people to move.
4. Eddie is very athletic.

5. Monica lost interest in playing baseball and wanted off the team.
6. He ugly.

7. The cat needs in.

8. I bought ﬁve pound of potatoes.

9. Peter needs off the plane.

10. ll ain’t going to do it.

1 1. The kids say it’s too hot outside, and they want in.

12. We go to the movies a lot anymore.

13. If you want in the safe, use this combination.

14. He don’t know nothing.

155

Read each of the following sentences to yourself. Rate each sentence on the following
scale:

15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.

34.

Choose a if you would never use this sentence.

Choose b if you would use this sentence only in writing or in very formal speech
situations like a job interview or in court. ,

Choose c if you would use this sentence only with close friends and/or family.

Choose d if you would use this sentence in general conversation, in stores, and with
people you don’t know well.

Choose e if you would use this sentence in all situations.

We need to get off at the next stop.

The sugar is all.

Dean needs in the building.

We had went to California on vacation.

This relationship is going nowhere. I want out.

He throwed the ball at the window.

The dog needs out three times a day.

Bill and me went to the game.

When the roller coaster picked up speed, Ed yelled that he wanted off.
He don’t know where she lives.

For his new job, Tony needs in the union.

He caught himself a cold.

The fumes from this bus are making me sick. I need off.
Where are you going on vacation?

Sam wants to get in the Army.

Ten o’clock is all the later I can stay up.

Do you want down?

Look at that dog come a-running.

I’m sorry to bother you, but I need by.

Mike says he wants to go off his medication.

156

Read each of the following sentences to yourself. Rate each sentence on the following
scale:

35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.

54.

Choose a if you would never use this sentence.

Choose b if you would use this sentence only in writing or in very formal speech
situations like a job interview or in court. ,

Choose c if you would use this sentence only with close friends and/or family.

Choose d if you would use this sentence in general conversation, in stores, and with
people you don’t know well.

Choose e if you would use this sentence in all situations.

Anymore, we eat ﬁsh often.

The day care center is full, and my child needs in.

We be tired.

When the cat meows at the door, she wants to go out.

Margaret seen him do it.

The baseball team hasn’t won a game in several weeks, and Jim wants off.
I’m not going to do it.

Barry’s job has become too stressful and he needs out.

Janet’s daughter is right smart.

This project is exactly what Donna likes to do. She wants in.

They are friendly.

Dan was getting nervous in the stuck elevator. He wanted out.
Jonathan bought dinner for Mike and 1.

I’ve wasted too much time working on this project, and I need off.
Bob’s a-ﬁxing the tractor.

This student wants in your class.

Do you want to come with?

In stormy weather, the plants on the porch need to come in.

He be singing all the time.

Randy is too busy and needs off the committee.

157

Read each of the following sentences to yourself. Rate each sentence on the following
scale:
Choose a if you would never use this sentence.
Choose b if you would use this sentence only in writing or in very formal speech
situations like a job interview or in court. ,
Choose c if you would use this sentence only with close friends and/or family.
Choose (1 if you would use this sentence in general conversation, in stores, and with
people you don’t know well.
Choose e if you would use this sentence in all situations.

55. I want you to ﬁx the chair right now.

56. Can you help the baby? She wants to get down.

 

57. They was singing while Jack was working.
58. I want off this horse.

59. I’m not going to paint the house this year.
60. The baby is on the couch and needs down.
61. Alice doesn’t eat meat anymore.

62. Mary took Bridget and me to school.

The following information is needed for statistical purposes. Please give only one
answer to each item.
63. Fill in l for this item. (This identiﬁes which questionnaire you completed)

64. a=Female b=Male

65. Age
a=18-24 years old b=25-34 years old c=35-54 years old
d=55-64 years old e=64 years old and over

66. With which racial/ethnic category do you most closely identify yourself?
a=European American (white) b=African American
c=Native American d=I-Iispanic =other

67. In the blank space on the op-scan form, please print the city and state where you
grew up (between the ages of 2 and 15).

68. In the blank space on the op-scan form, please print the profession of the primary
wage-earner in your family.

158

APPENDIX C

Essential information on Linguistic Atlas projects and DARE

 

159

Information presented here was gathered from the following sources:
Allen (1975); Cassidy (1985); Kurath (1939); Linguistic atlas projects
(www.hyde.park.uga.edu); McDavid & Payne (1976); Pederson (1991)

Table C. 1. Linguistic Atlas of New England (LANE)

 

States surveyed

Connecticut, Maine, Massachusetts, New Hampshire, New
York (Long Island), Rhode Island, and Vermont

 

Period of data collection

1931-1933

 

 

Number of respondents

 

213

 

Table C.2. Linguistic Atlas of the Middle and South Atlantic States (LAMSAS)

 

States surveyed

Delaware, northeastern Florida, eastern Georgia, Maryland,
New Jersey, New York, North Carolina, Pennsylvania,
South Carolina, Virginia, and West Virginia

 

Period of data collection

1933-1974

 

 

Number of respondents

 

1162

 

Table 03. Linguistic Atlas of the North Central States (IANCS)

 

States surveyed

Illinois, Indiana, Kentuclg, Michigan, Ohio, and Wisconsin

 

Period of data collection

1933-1978 ,

 

 

Number of respondents

 

564

 

Table 04. Linguistic Atlas of the Upper Midwest (LAUll/I)

 

States surveyed

Iowa, Minnesota, Nebraska, North Dakota, and South
Dakota

 

Period of data collection

1949-1962

 

 

Number of respondents

 

208

 

Table C.5. Linguistic Atlas of the Gulf States (LAGS)

 

: States surveyed

Alabama; Arkansas, Florida, Georgia, Louisiana,
Mississippig Tennessee, and east Texas

 

Period of data collection

1968-1983

 

 

Number of respondents

 

1121 (914 analyzed)

 

160

 

 

 

 

 

Table C .6. Linguistic Atlas of the Paciﬁc Northwest (LAPNW)

 

 

 

 

States surveyed Idaho, Oregon (few), Montana (few), and Washington
Period of data collection 1953-1963 '
Number of respondents 51

 

 

Table C.7. Dictionary of American Regional English (DARE)

 

States surveyed all

 

Period of data collection 1965-1970

 

 

 

Number of respondents Q02

 

161

 

 

APPENDIX D

List of respondents

162

Abbreviations:
AfA=African American
AI=American Indian
EA=European American
NR=No Response

Table D. 1. List of respondents

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

k Dialect Rural- SES

esp Hometown Region Urban Sex (Table 3.2) Age Race/Eth Quest
119 Blissﬁeld, MI North Rural M 4 18-24 EA A1
135 ewitt, MI North Urban M 1 18-24 Other B2
141 lDewitt, MI North Urban M 3 18-24 EA B2
133 ssexville, MI North Urban F 2 18-24 EA B1
106 Fannington Hills, MI North Urban M 5 1824 EA B2
132 Ferndale, MI North Urban F 2 18-24 AfA A2
143 Genesee, MI North Rural M 3 18-24 NR A1
111 Grand Rapids, MI North Urban M NR 18-24 EA A2
1 18 Kalamazoo, MI North Urban M 4 1 8-24 EA Bl
123 Kalamazoo, MI North Urban M 5 18-24 EA B2
122 lLakeview, MI North Rural M 4 18-24 EA A1
109 Lansing, MI North Urban F 1 18-24 EA B2
142 lansing, MI North Urban M 2 18-24 EA Bl
136 [Livonia, MI North Urban F 1 18-24 EA A1
116 [MarshalL MI North Rural M 6 18-24 EA A1
108 jvrnford, MI North Rural M 1 18-24 EA A2
134 [Mt Clemens, MI North Urban F 4 18-24 EA B]
114 lPetcrsbgg, MI North Rural M 4 1824 EA B1
144 oyal OakLMI North Urban M 1 18-24 EA A2
140 Saginaw, MI North Urban M 2 18-24 EA A2
110 Southﬁeld, MI North Urban M 4 1824 EA B2
113 Sterling Heights, MI North Urban M 2 18-24 EA A2
138 Sterling Heights, MI North Urban M 2 18-24 NR B2
107 Stevensville, MI North Rural F 4 25-34 EA A1
120 Troy, MI North Urban M 2 18-24 EA B1
121 Troy, MI North Urban M 3 18-24 EA A2
131 Warren, MI North Urban F 3 18-24 EA B1
145 WixorrL MI North Urban M 2 18-24 EA A2
112 Wyoming, MI North Urban M 4 18-24 EA B1
137 Ypsilanti, MI North Urban M NR 18-24 NR A2
117 Zeeland, MI North Urban F 2 18-24 Other Al
1 15 Canton, MI North Urban M 1 18-24 Other A]
124 Chicago, IL North Urban M 2 1 8-24 EA B l
1 Cleveland, OH North Urban M 3 18-24 EA B1
6 Cleveland, OH North Urban F 1 18-24 EA Bl

 

163

 

Table D. 1. (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dialect Rural- SES
Resp Hometown Region Urban Sex (Table 3.2) Age Race/Eth Quest
9 Cleveland, OH North Urban M 2 18-24 EA B1
16 Cleveland, OH North Urban M 3 18-24 EA B1
20 Cleveland, OH North Urban M 4 18-24 EA B1
60 Toledo, OH North Urban F NR 18-24 H A1
139 Toledo, OH North Urban M 4 18-24 EA A1
125 'Manawa, WI North Rural M 4 18-24 EA Al
14 North Canton, OH Midland Urban F 4 18-24 EA B2
23 Canton, OH Midland Urban M NR 18-24 EA B1
25 Galion, OH Midland Urban F 3 18-24 EA A2
4 [Mansﬁeld, OH Midland Urban F . 3 18-24 Other A2
31 [Mansﬁeld OH Midland Urban F 4 18-24 EA Al
81 'Mansﬁeld, OH Midland Urban M 4 18-24 EA B1
45 Coshocton, OH Midland Rural F 6 18-24 EA Al
50 Coshocton, OH Midland Rural M 3 25-34 EA B1
53 Coshocton, OH Midland Rural M 6 25-34 EA A2
64 Coshocton, OH Midland Rural F 2 18-24 EA A1
88 Coshocton, OH Midland Rural M 6 18-24 EA B2
22 Delaware, OH Midland Urban M 2 18-24 EA B2
67 Frederickstown, OH Midland Rural M 6 18-24 EA B2
80 [Mt VernonQH Midland Urban M 3 1 8-24 EA A l
91 jMt. Vernon, OH Midland Urban M 1 18-24 EA 132
62 New Philadelphia, OH Midland Urban F 4 18-24 EA Al
96 Utica, OH Midland Rural F 2 18-24 EA Al
90 Granville, OH Midland Urban M 6 18-24 Al A2
95 Granville, OH Midland Urban M 5 18-24 EA A2
44 Heath, OH Midland Urban M 6 18-24 EA B2
68 Heath, OH Midland Urban F 4 18-24 EA A2
70 Heath, OH Midland Urban F NR 18-24 EA BZ
73 Heath, OH Midland Urban M 4 18-24 EA B1
93 Heath, OH Midland Urban M 3 18-24 EA B1
32 Newark, OH Midland Urban M 2 35-54 EA B2
34 Newark, OH Midland Urban M 6 18-24 AfA Al
37 Newark, OH Midland Urban M 3 18-24 EA B2
42 Newark, OH Midland Urban M 2 18-24 EA B1
43 Newark, OH Midland Urban F 4 18-24 Other A2
46 Newark, OH Midland Urban F 4 18-24 EA Al
47 Newark, OH Midland Urban M NR 18-24 EA Al
48 NewarkLOH Midland Urban M 5 1 8-24 EA A2
49 Newark, OH Midland Urban M 6 25-34 AfA B1
57 Newark, OH Midland Urban M 4 18-24 EA B1
58 Newark, OH Midland Urban M 4 18-24 NR A1

 

164

 

Table D. 1. (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I; Dialect Rural- ' SES

esp ometown Region Urban Sex (Table 3.2) Age Race/Eth Quest
61 Newark, OH Midland Urban M 2 18-24 EA BZ
66 Newark, OH Midland Urban M 3 25-34 A1 B1
69 Newark, OH Midland Urban M 4 18-24 EA BZ
74 Newark, OH Midland Urban M 2 18-24 EA Al
76 Newark, OH Midland Urban F 5 18-24 EA A1
78 Newark, OH Midland Urban M 4 18-24 EA BI
83 Newark, OH Midland Urban M 3 18-24 EA B1
84 Newark, OH Midland Urban M 4 35-54 EA A1
85 Newark, OH Midland Urban M 4 18-24 EA BI
87 Newark, OH Midland Urban M 4 18-24 EA Al
89 Newark, OH Midland Urban F 5 18-24 AfA B1
92 Newark, OH Midland Urban F 4 18-24 EA B2
97 Newark, OH Midland Urban M 4 18-24 EA A2
98 Newark, OH Midland Urban F 1 18-24 EA B2
101 Newark, OH Midland Urban M 5 25-34 EA A2
102 ewark, OH Midland Urban F 6 18-24 EA A2
35 Johnstown, OH Midland Rural M 2 18-24 EA B2
99 Johnstown, OH Midland Rural M 6 35-54 EA A2
204 Johnstown, OH Midland Rural F 4 35-54 EA BI
59 Pataskala, OH Midland Rural M 5 18-24 EA A2
203 Pataskala, OH Midland Rural F 2 55+ EA A2
205 lPataskala, OH Midland Rural F 5 55+ EA A1
206 Pataskala, OH Midland Rural M 4 35-54 EA A2
207 Pataskala, OH Midland Rural M 5 55+ EA Bl
208 [Pataskala, OH Midland Rural F 5 35-54 BA 82
209 lPataskala, OH Midland Rural _ M 5 3554 EA A2
86 Reynoldsburg, OH Midland Urban M 3 18-24 EA A1
100 IBlacklick, OH Midland Urban M 2 25-34 EA Al
5 Columbus, OH Midland Urban M 3 18-24 Other 82
15 Columbus, OH Midland Urban M 3 18-24 EA B2
21 Columbus, OH Midland Urban M 2 55+ EA A2
24 Columbus, OH Midland Urban M NR 18-24 EA A1
52 Columbus, OH Midland Urban F 1 18-24 EA BZ
54 Columbus, OH Midland Urban F 5 18-24 EA Al
65 Columbus, OH Midland Urban M 5 18-24 Other B2
79 Columbus, OH Midland Urban M 4 18-24 EA B2
202 Columbus, OH Midland Urban M 3 55+ EA B2
3 Grove City, OH Midland . Urban F 6 18-24 EA A2
201 Hilliard, OH Midland Urban F 4 55+ EA A1
12 Upper Arlington, OH Midland Urban M 2 18-24 EA A2
63 Baltimore, OH Midland Rural M 3 25-34 EA A2

 

 

 

165

 

Table D. 1. (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IR Dialect Rural- SES
es Hometown Region Urban Sex (Table 3.2) Age Race/Eth Quest
56 lPickerington, OH Midland Rural M 5 25-34 EA B1
213 Pickerington, OH Midland Rural F 5 3554 EA A1
51 Pickerington, OH Midland Rural M 5 18-24 EA B2
94 Thomville, OH Midland Rural M 2 1 8-24 EA A 1
14 Gahanna, OH Midland Urban M 5 35-54 EA Bl
29 Pittsburgh, PA Midland Urban M 2 18-24 EA A2
77 Brazil, IN Midland Urban M 2 18-24 NR A2
1 1 ayton, OH Midland Urban F 4 18-24 EA Al
72 Roseville, OH Midland Rural M 5 25-34 EA BI
39 S. Bloomingville, OH Midland Rural M 4 18-24 EA B1
82 Summerﬁeld, OH Midland Rural ' F 4 35-54 BA 82
30 Xenia, OH Midland Urban M 3 18-24 EA B1
21 l Middleport OH Midland Rural F 5 55+ EA BZ
210 Wheelersburg, OH Midland Urban M 5 55+ EA A2
215 Wheelersburg, OH Midland Urban F 5 55+ EA A1
164 Wise, VA Midland Rural M 2 25-34 EA A2
212 Wise, VA Midland Rural M 5 55+ EA B1
2 8 Charleston, WV Midland Urban F 2 1 8-24 EA A 1
105 Weirton, WV Midland Urban M 2 25-34 EA B1
150 Atlanta, GA South Urban F 4 18-24 Other Bl
153 tlanta, GA South Urban M 2 18-24 EA A1
1 63 AtlantaLGA South Urban M 3 1 8-24 EA A 1
167 Atlanta, GA South Urban F 2 25-34 NR A1
172 Atlanta, GA South Urban M 2 18-24 EA A2
17 0 Cohutta, GA South Rural M 2 18-24 EA B2
161 Dunwoody, GA South Urban F 3 18-24 EA A1
157 J onesboro, GA South Rural M 3 18-24 EA Bl
174 Lawrenceville, GA South Urban M 4 25-34 EA B2
159 Loganville, GA South Rural M 5 18-24 EA B2
175 Macon, GA South Urban M 3 18-24 EA A2
176 Macon, GA South Urban F 3 18-24 AfA A l
152 'Marietta, GA South Urban M 5 18-24 EA A1
156 Marietta, GA South Urban M 2 18-24 EA A2
166 Marietta, GA South Urban M 3 18-24 EA Bl
168 Martinez, GA South Urban M 2 18-24 EA A2
171 Savannah, GA South Urban M 3 18-24 EA Bl
155 Stone Mountain, GA South Urban M 4 18-24 EA A2
160 Stone Mountairg GA South Urban M 3 18-24 Other A2
1 54 Suwanee, GA South Rural F 4 1 8-24 EA A 1
149 Thomasville, GA South Urban F 2 18-24 AfA BZ
165 aldosta, GA South Urban F l 1 8-24 EA A2

 

 

 

 

 

 

 

 

 

 

166

 

Table D. l. (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Dialect Rural- SES
Resp ometown Region Urban Sex (Table 3.2) Age Race/Eth Quest
151 Wintcrville, GA South Rural M 2 25-34 NR A1
162 Woodstock, GA South Rural F 2 18-24 EA B2
173 Myrtle Beach, SC South Urban M 1 18-24 EA A2
33 Nofolk, VA South Urban ‘ M 4 18—24 AfA A2
169 Nofolk, VA South Urban M 4 18-24 EA B1

 

167

 

APPENDIX E

Learning curve statistics

168

Table E.1. Use of token sentences in three groupings corresponding to the beginning,

middle, and end of the questionnaire

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Version N per First 3 Second 8 Third 8 2 Pearson
sector tokens tokens tokens x ldfl 9
North
A” 3 28 (5168?) (5129?) (5168?) 0.45 2 .800
2.2 88 322/; 23;; :3; 2
A2 72 3:; 2:3 2:3; 1.82 2 .403
B] “2 258/; 23:; (233/; 7.19 2 .027
° 0 o
22 22 {2’23" 2:; 3,7,; 2
Midland
0 0 0
All 760 (7599?) (is?) (:22?) 5.73 2 .057
81 82°33 5%; (2126.32) 2
A2 ’92 (718$) 51%?) (8166?) 4.81 2 .090
B1 176 (6151?) (712:3 (7153?) 8.19 2 .016
22 552?. 597?, 5,86%, 2
South
A“ 2 1 6 (5131‘?) (ii?) 8%) 2.65 2 .266
81 84 £33; :33; 23:; 2 .222
82 72 222/; 21:; :33; 2
° 0 0
BI 40 (2%, 383° (2537/; 1.26 2 .531
° 0 0
B2 40 32%) 2%; 232/; 5.21 2 .074

 

 

 

 

 

 

 

 

169

 

REFERENCES

Adams, Michael. 2000. Lexical doppelgangers. Journal of English Linguistics 28. 295-
310.

Albanyan, Ahmed, and Dennis R. Preston. 1998. What is standard American English?
Studia Anglica Posnaniensia 33.29-46.

Allen, Harold B. 1975. The linguistic atlas of the Upper Midwest. Vol. 2. Minneapolis:
University of Minnesota Press.

The American Heritage college dictionary, 4th ed. 2002. Boston: Houghton Miﬁlin Co.

Ashcom, B. B. 1953. Notes on the language of the Bedford, Pennsylvania, subarea.
American Speech 28241-25529.

Asher, Joseph. 2001. Serving the wealthy: Everybody wants in. ABA Banking Journal
93 .42-44.

Aurand, Jr. A. Monroe. 1939. Quaint idioms and expressions of the Pennsylvania-
Germans. Harrisburg, Pennsylvania: The Aurand Press.

Bailey, Charles-James N. 1968. Is there a Midland dialect of American English? ERIC
Document 021 240.

Bailey, Guy, and Jan Tillery. 1999. The Rutledge effect: The impact of interviewers on
survey results in linguistics. American Speech 74.3 89-402.

Bailey, Guy, Tom Wikle, and Jan Tillery. 1997. The effects of methods on results in
dialectology. English World-Wide 18.35-63.

Bailey, Guy, Tom Wikle, Jan Tillery, and Lori Sand. 1991. The apparent time construct.
Language Variation and Change 3.241-264.

----- . 1993. Some patterns of linguistic diffusion. Language Variation and Change 5.3 59-
390.

Bell, Allan 1984. Language style as audience design. Language in Society 13.145-204.

Benson, Erica J. to appear. Folk linguistic perceptions and the mapping of dialect
boundaries. American Speech.

170

Bright, Elizabeth S. 1971. A word geography of California and Nevada. University of
California Publications in Linguistics 69. Berkeley, California: University of
California Press.

Bryant, Margaret M. 1962. Current American usage. New York: Funk & Wagnalls.

Burnham, Josephine M. 1926. Some observations upon Middle Western speech. Dialect
Notes 5.391-396.

Carr, J. W. 1904. A list of words from Northwest Arkansas. Dialect Notes 2.416-422.

Carver, Craig. 1987. American regional dialects. Ann Arbor: University of Michigan
Press.

Carlton, Esq, Robert (pseudonym for Baynard Rush Hall). 1843. The new purchase: or
seven and a half years in the Far West. Vol. I. New York: D. Appleton and Co.

----- . 1916. The new purchase: or seven and a half years in the Far West, 3rd edition
(Indiana Centennial edition). Princeton, New Jersey: Princeton University Press.

Carruth, W. H. 1892. Dialect word-list—No. 2. Kansas University Quarterly 1.137-142.

Cassidy, Frederic G. (ed.) 1985. Dictionary of American regional English, vol. 1.
Cambridge, Massachusetts: The Belknap Press of Harvard University Press.

Cassidy, Frederic G., and Joan Houston Hall (eds) 1996. Dictionary of American

regional English, vol. III: I-O. Cambridge, Massachusetts: The Belknap Press of
Harvard University Press.

----- to appear. Dictionary of American regional English, vol. V: Sl-Z. Cambridge,
Massachusetts: The Belknap Press of Harvard University Press.

Chambers, J.K. 1995. Sociolinguistic theory: linguistic variation and its social
signiﬁcance. Oxford: Blackwell.

Chambers, J .K., and Peter Trudgill. 1980. Dialectology. Cambridge: Cambridge
University Press.

----- 1991. Dialect grammar: Data and theory. Dialects of English: Studies in grammatical
variation, ed. by Peter Trudgill, and J. K. Chambers, 291-296. London: Longman.

Crozier, Alan. 1984. The Scotch-Irish inﬂuence on American English. American Speech
59.310-331.

Crumb, D.S. 1903. The dialect of Southeast Missouri. Dialect Notes 2.304-337.

171

Dakin, Robert Ford. 1966a. The dialect vocabulary of the Ohio River Valley, Vol. 1. Ann
Arbor, Michigan: University of Michigan Dissertation.

Dakin, Robert Ford. 1966b. The dialect vocabulary of the Ohio River Valley, Vol. 11.
Ann Arbor, Michigan: University of Michigan Dissertation.

Davis, Lawrence M., and Charles L. Houck. 1992. Is there a Midland dialect area?
—Again. American Speech 67.61-70.

----- . 1996. The comparability of Linguistic Atlas records: The case of LANCS and
LAGS. Focus on the USA, ed. by Edgar W. Schneider, 51-62. Amsterdam: John
Benjamins.

DeRamus, Betty. 2002. Want off the darn spam e-mail list? You’re better off deleting the
junk. The Detroit News, January 15, 2002.
www.detnews.com/2002/metro/0201/15/b01-390701.htm.

Downes, William. 1998. Language and society. Cambridge: Cambridge University Press.

Eckert, Penelope. 1989. The whole woman: Sex and gender differences in variation.
Language Variation and Change 1.245-267.

----- 1991. Social polarization and the choice of linguistic variants. New Ways of
Analyzing Sound Change, ed. by Penelope Eckert, 213-232. New York:
Academic Press.

----- 1997. Age as a sociolinguistic variable. The handbook of sociolinguistics, ed. by
Florian Coulmas, 151-167. Oxford: Blackwell.

Eckert, Penelope, and Sally McConnell-Ginet. 1992. Think practically and look locally:
Language and gender as community-based practice. Annual Review of
Anthropology 21 .461-490.

Eisikovits, Edina. 1989. Girl-talk/boy-talk: Sex differences in adolescent speech.
Australian English: The language of a new society, ed. by Peter Collins and David
Blair, 35-54. Queensland: University of Queensland Press.

Eliason, Norman Ellsworth.1956. Tarheel talk: an historical account of the English
language in North Carolina to 1860. Chapel Hill, North Carolina: University of
North Carolina Press.

Evans, Bergen, and Cornelia Evans. 1957. A dictionary of contemporary American
usage. New York: Random House.

172

Finegan, Edward, and Douglas Biber. 1994. Register and social dialect variation: an
integrated approach. Sociolinguistic perspectives on register, ed. by Douglas
Biber and Edward Finegan, 315-347. New York: Oxford University Press.

Flanigan, Beverly Olson. 2000. Mapping the North Centralstates: Changing views.
Powerpoint poster at NWAV 29.

Flanigan, Beverly Olson, and Franklin Paul Norris. 2000. Cross-dialectal comprehension
as evidence for boundary mapping: Perceptions of the speech of southeastern
Ohio. Language Variation and Change 12.175-201.

Foley, Lawrence M. 1972. A phonological and lexical study of the speech of Tuscaloosa
County, Alabama. Publication of the American Dialect Society 38. Tuscaloosa,
Alabama: University of Alabama Press.

Frazer, Timothy C. 1982. Joseph Kirklands’ Zuiy as linguistic evidence. American
Speech 57.190-197.

----- 1987. Attitudes toward regional pronunciation. Journal of English Linguistics 20.89-
100.

----- 1993. Problems in Midwest English: Introduction and overview. “Heartland”
English, ed. by Timothy C. Frazer, 1-19. Tuscaloosa, Alabama: The University of
Alabama Press.

----- 1994. On transition areas and the “Midland” dialect: A reply to Davis and Houck.
American Speech 69.430-435.

----- 1996. The dialects of the Middle West. Focus on the USA, ed. by Edgar W.
Schneider, 81-102. Amsterdam: John Benj amins.

Frazer, Timothy C ., Thomas E. Murray, and Beth Lee Simon. 1996. Need + past
participle in American English. American Speech 71.255-271.

Fridland, Valerie Michelle. 1998. The Southern Vowel Shift: Linguistic and social
factors. East Lansing, Michigan: Michigan State University Dissertation.

Goodridge, Elisabeth. 2001. Kodak wants out of the classroom. Information Week
831.94, April 2, 2001.

Greenbaum, Sidney, and Janet Whitcut. 1988. Longman guide to English usage. Essex:
Longman.

Griebling, Erich T. 1947. Regional locutions. Word Study 22.5.

173

Hall, Joan Houston. 2002. Personal communication: Materials from DARE archives and
project data.

Hankey, Clyde T. 1960. A Colorado word geography. Publication of the American
Dialect Society 34. University, Alabama: University of Alabama Press.

-----1961. Semantic features and Eastern relics in Colorado dialect. American Speech
36.266—270.

Hartman, James W. 1966. Pressures for dialect change in Hocking County, Ohio. Ann
Arbor, Michigan: University of Michigan Dissertation.

Harvey, Bartle T. 1914. Addenda to the word-list from the Northwest. Dialect Notes
4.162-164.

Haynes, John. 1995. Style. London: Routledge.

Heydrick, B. A. 1908. Provincialisms of Southeastern Pennsylvania. German American
Annals 10.32-52.

Hodge, Robert W., Paul M. Siegel, and Peter H. Rossi. 1964. Occupational prestige in the
United States, 1925-63. American Journal of Sociology 70.286-302.

Hudson, Grover. 2003. Personal communication: Discussion of an earlier draft of this
dissertation.

Irvine, Judith T. 2001. “Style” as distinctiveness: The culture and ideology of linguistic
differentiation. Style and sociolingiustic variation, ed. by Penelope Eckert and
John R. Rickford, 21-43. Cambridge: Cambridge University Press.

Ito, Rika. 1999. Diffusion of urban sound change in rural Michigan: A case of the
Northern Cities Shift. East Lansing, Michigan: Michigan State University
Dissertation.

Jacobson, Hannah. 1931. Milwaukee dialect. Madison, Wisconsin: University of
Wisconsin, BA Thesis.

Johnson, Ellen. 1994. Yet again: The Midland dialect. American Speech 69.419-430.
Joos, Martin. 1967. The ﬁve clocks. New York: Harcourt, Brace & World, Inc.

Karpinski, Richard. 2002. IT wants off the Windows upgrade merry-go-round.
Internetweek 890.PG15. January 7, 2002.

Kirkland, Joseph. 1887. Zury: The meanest man in Spring County: A novel of Western
life. Boston: Houghton Miﬁlin and Company.

174

Krapp, George Philip. 1925. The English language in America. New York: Frederick
Ungar Publishing Company.

Krumpelmann, John T. 1939. West Virginia Peculialities. American Speech 14.155-156.

Kurath, Hans (ed.). 193 9. Handbook of the linguistic geography of New England.
Providence, Rhode Island: Brown University.

----- 1943. Linguistic atlas of New England, Maps 63-734. Providence, Rhode Island:
Brown University.

-----1949. A word geography of the Eastern United States. Ann Arbor: University of
Michigan Press.

Labov, William. 1964. Stages in the acquisition of standard English. Social dialects and
language learning, ed. by Roger W. Shuy et al, 77-103. Champaign, Illinois:
National Council of Teachers of English.

-----1966. The social stratiﬁcation of English in New York City. Washington DC: Center
for Applied Linguistics.

----- 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press.

----- 1973. Where do grammars stop? Monograph on Languages and Linguistics 25, ed.
by Roger W. Shuy, 43-88. Washington, DC: Georgetown University Press.

-----1975. Empirical foundations of linguistic theory. The scope of American linguistics,
ed. by Robert Austerlitz, 77-133. Lisse, Belgium: The Peter de Ridder Press.

----- 1978. Where does the linguistic variable stop? A response to Beatriz Lavandera.
Working Papers in Sociolinguistics 44.

----- 1990. The intersection of sex and social class in the course of linguistic change.
Language Variation and Change 2.205-254.

----- 1991. The boundaries of a grammar: Inter-dialectal reactions to positive anymore.
Dialects of English: Studies in grammatical variation ed. by Peter Trudgill and J.
K. Chambers, 273-288. London: Longman.

----- 2001. Principles of linguistic change. Volume 2: Social factors. Malden,
Massachusetts: Blackwell.

Labov, William, Sharon Ash, and Charles Boberg. 1997. A national map of the regional

dialects of American English. TELSUR project, Department of Linguistics,
University of Pennsylvania, http://www.ling.upenn.edu/phono_atlas.

175

----- to appear. Atlas of North American English. Berlin: Mouton de Gruyter.

Lance, Donald M. 1994. “Variation in American English.” In American Pronunciation
(12th edition, expanded), John Kenyon, 345-501. Ann Arbor: George Wahr.

Langacker, Ronald W. 1999. Losing control: Grammaticization, subjectﬁcation, and
transparency. Historical semantics and cognition (Cognitive Linguistics Research
13), ed. by Andreas Blank and Peter Koch, 147-175. Berlin: Mouton de Gruyter.

Lavandera, Beatriz R. 1977. Where does the sociolinguistic variable stop? Working
Papers in Sociolinguistics 40.

-----1988. The study of language in its socio-cultural context. Linguistics: The Cambridge
survey, Vol. IV: The sociocultural context, ed. by Frederick J. Newmeyer, 1-13.
Cambridge: Cambridge University Press.

Lehman, B. H. 1921. A Word List from California. Dialect Notes V. 109-1 14.

Leonard, Sterling Andrus. 1932. Current English usage. Chicago, Illinois: The Inland
Press. Publication of the National Council of Teachers of English.

Linguistic atlas of the Gulf states concordance. httpz/lwww.hyde.park.uga.edu/lags/.
University of Georgia.

Linguistic atlas projects. http://www.hyde.park.uga.edu. University of Georgia.

Lt'ldtke, Helmut. 1999. Diachronic semantics: Towards a uniﬁed theory of language
change? Historical semantics and cognition (Cognitive Linguistics Research 13),
ed. by Andreas Blank and Peter Koch, 49-60. Berlin: Mouton de Gruyter.

Marckwardt, Albert H. 1948. Want with ellipsis of verbs of motion. American Speech
23.3-9.

Marckwardt, Albert H., and Fred. G. Walcott. 1938. Facts about current English usage.
New York: Appleton-Century Crofts. Publication of the National Council of
Teachers of English.

Maxﬁeld, E. K. 1931. The speech of South-Western Pennsylvania. American Speech
7.18-20.

McAtee, W.L. 1942. Rural dialect of Grant County, Indiana, in the ‘nineties. Printed
privately.

----- 1956. Some dialect of Randolph County and elsewhere in North Carolina. Chapel
Hill, NC: Privately printed.

176

McCrum, Robert, William Cran, and Robert MacNeil. 1986. The story of English. New
York: Viking.

McDavid, Jr., Raven I. 1958. The dialects of American English. The structure of
American English, ed. by W. Nelson Frances, 480-543. New York: The Ronal
Press Co.

-----1964. Postvocalic —r in South Carolina: A social analysis. Language in culture and
society, ed. by Dell Hymes, 473-482. New York: Harper & Row.

----- 1983. The failure of intuition. The SECOL Review 7 . 128-139.

McDavid, Jr., Raven I., and Richard C. Payne (eds). 1976. Linguistic atlas of the North-
Central states: Basic materials 8. Microﬁlm collection of manuscripts on cultural
anthropology. Series 35. Chicago, Illinois: Joseph Regenstein Library, University
of Chicago.

McElhinny, Bonnie. 1999. More on the third dialect of English: Linguistic constraints on
the use of three phonological variables in Pittsburgh. Language Variation and
Change 11.171-195.

Merriam-Webster’s collegiate dictionary. 2003. On-line edition. Merriam-Webster.
www.m-w.com/netdict.htm.

Modaressi-Tehrani, Y. 1978. A sociolinguistic analysis of Modern Persian. Lawrence,
Kansas: University of Kansas Dissertation.

Montgomery, Michael. 1991. The roots of Appalachian English: Scots-Irish or British
Southern? Southern Appalachia and the South: A region within a region, ed. by
John C. Oscoe, 177-191. Special issue of the Journal of the Appalachian Studies
Association 3.

----- 1997 . The Scotch-Irish element in Appalachian English: How broad? How deep?
Ulster and North America: Transatlantic perspectives on the Scotch-Irish, ed. by
Curtis Wood and Tyler Blethen, 189-212. Tuscaloosa, Alabama: University of
Alabama Press.

Morris, William, and Mary Morris. 1985. Harper dictionary of contemporary usage. New
York: Harper Perennial.

Mullen, Kate. 1925. Westemisms. American Speech 1.149-153.

Murray, Thomas E. 1993. Positive anymore in the Midwest. “Heartland” English, ed. by
Timothy C. Frazer, 173-186. Tuscaloosa, Alabama: University of Alabama Press.

177

----- 2002. Language variation and change in the urban Midwest: The case of St. Louis,
Missouri. Language Variation and Change 14.347-361.

Murray, Thomas E., and Beth Lee Simon. 1999. Want + past participle in American
English. American Speech 74.140-164.

-----2002. Like + past participle in American English. American Speech 77.32-69.

Nakao, Keiko, and Judith Treas. 1994. Updating occupational prestige and
socioeconomic scores: How the new measures measure up. Sociological
Methodology 24.1-72.

Noguera, Pedro A. 2000. When parents want out. The Unesco Courier 11.24-26.

Norman, Arthur M.Z. 1971. A Southeast Texas dialect study. Readings in American
dialectology, ed. by Harold B. Allen and Gary N. Underwood, 135-151. New
York: Appleton-Century Crofts.

Oxford American dictionary and language guide. 1999. New York: Oxford University
Press.

Oxford English Dictionary. 1987. 2”d edition. Oxford: Clarendon Press.

Pederson, Lee (ed.). 1991. Linguistic atlas of the Gulf states. Athens, Georgia: University
of Georgia.

Pederson, Lee, Raven I. McDavid, Jr., Charles W. Foster, and Charles E. Billiard (eds)
1974. A manual of dialect research in the Southern states. University, Alabama:
The University of Alabama Press.

Pollard, Mary 0. 1915. Terms from the Tennessee mountains. Dialect Notes 4.242-243.

Pound, Louise. 1905. Dialect speech in Nebraska. Dialect Notes 3.55-67.

Preston, Dennis R 1991. Sorting out the variables in sociolinguistics theory. American
Speech 66.33-56.

Prettyman, William. 1907. Dialectal peculiarities in the Carlisle vernacular. German
American Annals V (Old Series IX).67-79.

Quirk, Randolph, Sidney Greenbaum, Geoffrey Leech, and Jan Svartvik. 1985. A
comprehensive grammar of the English language. London: Longman.

The Random House dictionary of the English language. 1987. 2"d edition. Unabridged.
New York: Random House.

178

Rice, William. 1902. The pioneer dialect of Southern Illinois. Dialect Notes 2.224-249.

Romaine, Susan. 1980. Stylistic variation and evaluative reactions to speech. Language
and Speech 23.213-232.

Russell, 1. Willis. 193 9. Prescription and description in English language usage.
American Speech 14.291-296.

Shaw, Harry. 1975. Dictionary of problem words and expressions. New York: McGraw-
Hill.

Shuy, Roger. 1962. The Northern-Midland dialect boundary in Illinois. Publication of
the American Dialect Society 38. Tuscaloosa, Alabama: University of Alabama
Press.

Simon, Beth Lee and Thomas E. Murray. 1999. How suite it is: Using combined
methodologies to explore dialect questions. Journal of English Linguistics 27.27-
39.

Spears, Richard A. 1987. American idioms dictionary. Lincolnwood, Illinois: National
Textbook Co.

Stanley, Oma. 1936b. The speech of East Texas. Chapter 4. American Speech 11.27-355.

Thomas, Erik. R. 1993. The use of all the + comparative structure. “Heartland” English,
ed. by Timothy C. Frazer, 257-265. Tuscaloosa: University of Alabama Press.

Traugott, Elizabeth Closs. 1982. From prepositional to textual and expressive meanings:

Some semantic-pragmatic aspects of grammaticalization. Perspectives on
historical linguistics, ed. by Winfred P. Lehmann and Yakov Malkiel, 245-271.
Amsterdam: John Benjamins.

-----1995. Subjectiﬁcation in grammaticalization. Subjectivity and subjectivisation:
Linguistic perspectives, ed. by Dieter Stein and Susan Wright, 31-54. Cambridge:
Cambridge University Press.

Trudgill, Peter. 1972. Sex, covert prestige and linguistic change in urban British English.
Language in Soceity 1.179-195.

----- 1974. The social differentiation of English in Norwich. Cambridge: Cambridge
University Press.

----- 2002. Sociolinguistic variation and change. Washington, DC: Georgetown University
Press.

Warner, W. Lloyd. 1960. Social class in America. New York: Harper.

179

Warnick, Florence. 1942. Dialect of Garrett County, Maryland. Printed privately.

Webster’s third new international dictionary of the English language. 1961. Unabridged.
Springﬁeld, Massachusetts: G & C Merriam Co.

Wentworth, Harold 1944. American dialect dictionary. New York: Thomas Y. Crowell
Co.

Williams, Joseph M. 1975. Origins of the English language: A social and linguistic
history. New York: The Free Press.

Wilson, Arthur. 1948. English spoken by Pennsylvania Germans in Snyder County,
Pennsylvania. American Speech 23.23 6-23 8.

Wodak, Ruth, and Getraud Benke. 1997. New perspectives on variation studies. Gender

as a sociolinguistics variable. The handbook of sociolinguistics, ed. by Florian
Coulmas, 127-150. Oxford: Blackwell.

Wilson, Kenneth G. 1993. The Columbia guide to standard American English. New
York: Columbia University Press. www.bartlebycom/68/.

Wolfram, Walt and Donna Christian. 1976. Appalachian speech. Arlington, Virginia:
Center for Applied Linguistics.

Wolfram, Walt, and Ralph Fasold. 1974. The study of social dialects in American
English. Englewood Cliffs, New Jersey: Prentice-Hall, Inc.

Wolfram, Walt, and Natalie Schilling-Estes. 1998. American English: Dialects and
variation. Oxford: Blackwell Publishers.

Youmans, Gilbert. 1986. Any more on anymore?: Evidence from a Missouri dialect
survey. American Speech 61 .61-75.

Zagorin, Adam. 2000. Why Libya wants in. Time 155.66-67, March 27, 2000.

180

11