 

 

”:‘i

3:5:

v
"r
A...»

~
0.

'0

3'3.-

«-
-a

..._
m

I a.

 

 

THESIS

LIBRARIES

Illlllllllllllllllllllllllllllllllllll III |

3 1293 01688 75

 

 

 

 

 

 

 

 

This is to certify that the I

thesis entitled

A unified analysis of negative inversion ]

constructions in African American English

presented by

Walter Lee Sistrunk, Jr.

has been accepted towards fulﬁllment
of the requirements for

Master's degree in LiﬂﬁUiStiCS

 

 

MAL/L ‘,

Major professor

 

 

Date Mai 10, 1998

 

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

 

 

LIBRARY

Mlchlgan State
Unlverslty

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

MTE DUE DATE DUE DATE DUE

 

:«E‘Zé E31

 

 

 

 

92:03 M

l—

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IM WWW“

A Uniﬁed Analysis of Negative Inversion sentences in African American English
By
Walter Lee Sistrunk Jr.

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

MASTERS OF ARTS
Department of Linguistics and Languages
1998

ABSTRACT

A UNIFIED ANALYSIS OF NEGATIVE INVERSION CONSTRUCTION S IN
AFRICAN AMERICAN ENGLISH

By
Walter Lee Sistrunk Jr.

This thesis provides a uniﬁed analysis of Negative Inversion (NI) constructions in
African American English. Earlier analyses done by Labov (1972) Sells, Rickford ,and
Wasom, (1996), and Weldon (1997) argue that two analyses are needed because expletive
subjects are not allowed in sentneces containing a madol auxiliary. It is shown that all NI
sentences must adhere to the defmiteness restriction, and therefore, the subjects must be
VP-intemal in all NI sentences. Lasnik’s (1995) argument is ultilized to show that
Negative Inversion sentences are like the existential sentences of Standard English in that
both the expletive subject and its associate need to be assigned Case independently. I argue
that expletives cannot appear in certain Negative Inversion sentences in African American
English because of Case requirements. Modal auxiliaries, unlike the copula verb be
cannot assigne partitive Case. In these cases, the associate must raise at Lf to be assigned
Case, and the expletive cannot appear. This allows all NI constructions to be given a

uniﬁed analysis which does not invlove inversion at all.

Copyright by
WALTER LEE SISTRUNK JR.
1998

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my Lord and Saviour for giving me the
strength to complete this work. A special thank my committee members for giving praise
and criticism that was both balanced and blunt. I would also like to thank my chair, Alan
Munn, whose support ranged from personal academic advice to ﬁling my tax extention
when the fear of my April 16th thesis defense was greater than the fear of federal
imprisonment from having failed to ﬁle my income tax. Last but not least, to my loving
family and network of friends who taught me to be a true "rhino," and that true happiness

comes from overcoming hardships.

iv

TABLE OF CONTENTS

INTRODUCTION .................................................................................... 1
CHAPTER 1 .......................................................................................... 3
1.1 Labov’s analysis of Negative Inversion ................................................ 3

1.2 Martin’s analysis of negative inversion sentence ...................................... 6

1.3 An Optimality account of NI ........................................................... 17

1.4 Weldon’s account on N I sentences .................................................... 23
1.4.1 Existential Quantiﬁers .......................................................... 23

1.4.2 The Non-copular Case ......................................................... 24

1.4.3 The Copular be2 Verb ......................................................... 25

1.4.4 Ambiguous Cases .............................................................. 28

1.4.5 Existential Cases ................................................................ 29

1.5 Summary ................................................................................. 31
CHAPTER 2 ........................................................................................ 33
2.1 Problems with NI setences as Fonted Negative Constructions .................... 33
2.2 Problems with Sells et al. 1996 ........................................................ 39

2. 3 Problems with Martin’s Deﬁnitenes Effect and Null Expletive account. .......... 41
2.4 Problems with Weldon’s HPSG account ............................................. 42
2.5 Summary ................................................................................. 50
CHAPTER 3 ........................................................................................ 52
3.1 Case Theory .............................................................................. 5 3
3.2 Case and Expletives ..................................................................... 59
3.3 Case and NI constructions .............................................................. 61
3.4 Individual Level predicates in NI sentences .......................................... 68
3.5 Summary ................................................................................. 69
CONCLUSION ..................................................................................... 71
REFERENCES ..................................................................................... 74

INTRODUCTION

Negative Inversion (NI) constructions in African American English (AAE) have
been studied by linguists for over thirty years. NI sentences are double negative
constructions having an initial negative auxiliary followed by a negative indeﬁnite subject.
Though these constructions look like interrogatives with an inverted auxiliary, they are
declarative sentences spoken with a rising intonation instead of the falling intonation that is
typical of interrogative sentences. They can have either a form of the verb be or a modal

verb as the initial auxiliary as shown below.

(1)
a. Ain’t nobody in their right mind leavin’ out in this bad weather

b. Won’t nobody go to the party tonight.

In this thesis, I will provide a uniﬁed analysis of both kinds of sentences.
In chapter one, I will provide an overview of the literature on Negative Inversion in African
American English. This overview will summarize several competing theories which fall
into two camps. The ﬁrst camp argues that two analyses are needed; one to account for the
N I construction with a modal auxiliary and a different one with the verb be being
existential. They argue that two structurally different negative expressions can be formed
depending on what type of auxiliary appears initially. Thus, negative constructions that
use existential auxiliaries are given an existential analysis, while on the other hand, negative
constructions that use a modal auxiliary are given an inversion analysis.

The argument for two analyses is based on the premise that in sentences where the
negative auxiliary is a form of the verb be, the expletive there can appear grammatically,

and an existential analysis is given. On the contrary, in a sentence having a modal as its

negative auxiliary, if the expletive appears the sentence is ungrammatical. Therefore,
sentences with modal auxiliary are given an inversion analysis.

The second camp argues that there is only one analysis needed to account for the
negative inversion construction. This argument is based on the fact that both the NI
construction made with the existential auxiliary and the modal auxiliary place special
conditions on its subject. Like existential sentences, this condition requires subject to be
indefinite. This condition is called the deﬁniteness restriction.

In chapter 2, I will give a criticism of each analysis by ﬁrst showing that the
defmiteness restriction is not accounted for by those who argue for two analyses are
needed. At the same time, I will show that those who argue for a single existential analysis
cannot account for why an overt expletive cannot appear in NI sentences that have modal
auxiliaries.

In chapter 3, I will argue that all N I constructions can be accounted for under an
existential analysis. An existential analysis will account for why the deﬁniteness effect is
required of the post auxiliary subject. I will then use Cas theory to explain why the

expletive must not appear in sentences containing modal auxiliaries.

CHAPTER 1

1.1 Labov’s analysis of Negative Inversion

Labov (1972) gives one of the earliest analyses of Negative Inversion (NI)
constructions. In this analysis, he states that negative inversion is a device which is used
as an option to add emphasis to negative sentences with indeterminate subjects. The
negative feature is placed on the auxiliary through the process of negative transfer. The
negative feature is transferred from the subject onto the verb. After negative transfer has
taken place the auxiliary verb undergoes the process of inversion.

(2)
a. Nobody will catch us—>
b. Nobody won’t catch us—>

c. Won’t nobody catch us—-)

(Labov 1972, pg, 811)

Labov states that the process of inversion is a component of negative attraction
involving the attraction of the negative features which transform the indeterminate any,
ever, and either into their negative alternates no, never and neither. The conditions
for negative attraction require that the negative feature attach to the ﬁrst indeterminate; this
attachment is obligatory if the subject is an indeterminate. Labov illustrates how negative
attraction is optional, but obligatory when an indeterminate subject is present. When the

indeterminate subject is skipped, the sentence is ungrammatical.

(3)
a. * He didn’t know anything about anybody. (AAE)
b. * He ever know nothing about anybody. (AAE)
c. * He knew nothing about anybody. (AAE)
d. * He knew anything about nobody. (AAE)
(Labov 1972, pg. 776)

In (3a), the auxiliary contains the negative feature but the indeterminate subject does
not. Here, the negative feature is not transferred from auxiliary to the indeterminate
subject. To account for (3b)’s ungrammaticality, the indeterminate ever, being the first
indeterminate lacks the negative feature, therefore, the negative feature cannot be assigned
to the indeterminate subject. In (3c), which minimally contrasts with (3b), the object is not
assigned the negative feature is not satisfying the negative transfer rule. In (3d), the ﬁrst
indeterminate is skipped and is not assigned the negative feature, which violates negative
transfer.

Labov (1972) contends that within the NI construction there are two sentence
structures: the null expletive structure and the negative inversion structure. Sentences that
begin with a negative auxiliary having some form of the copula-be are analyzed as
existential constructions containing a null expletive there. On other hand, sentences that
do not begin with a copula-be auxiliary do not contain null expletives and are analyzed as

inversion constructions.

(4)

a. Ain’t nobody in the room. (existential)
b. Ain’t nothing going on here.

c. Won’t nobody stop us. (inversion)

(I. Can’t nobody do that there.

Labov illustrates that adding an expletive to a N1 construction will produce either a
grammatical or an ungrammatical sentence. In sentences where the expletive forms a
grammatical construction, the sentence is an existential construction. Conversely, when
adding an overt expletive forms an ungrammatical construction, the sentence is an inversion

construction. This is shown in (5).

(5)
a. There/it ain’t nobody in the room.

b. * There/it won’t nobody stop us.

Labov argues that the inversion sentence is derived by a movement operation called
negative inversion (NI). The movement operation of NI involves the movement of the
negative auxiliary verb from the right of the subject to the left of the subject. Labov argues
that this movement operation is similar to the movement operation that occurs in Fronted

Negative Sentences in Standard English (SE) as shown in (7b).

(6)
a. Nobody can’t stop us.

b. Can’t nobody stop us.

(7)
a. John will never again agree to go on a blind date.

b. Never again will John agree to go on a blind date.

The examples above illustrate Labov’s argument: the auxiliary in (6a) and (7a) start
on the right of the subject. In (6b) and (7b) the auxiliary (along with or as a result of the
movement of the negative item) moves to the left of the subject . In SE, a negative
Adverbial must precede the fronted auxiliary.

In summary, Labov states that NI sentence constructions have two structures, the
existential type and the inversion type. The latter sentence type, which contain modal
auxiliaries like won’t, are said to have derived from an underlying sentence where the verb
has moved from the position where it originated, that being to the right of the subject.
Conversely, Labov holds that in sentence where the negative auxiliary is in some form of
the copula be such as ain’t , there is a possibility that these sentences are existential

sentences with a deleted expletive.

1 .2 Martin’s analysis of negative inversion sentences

Martin (1992) presents an argument that NI constructions are not inversion
constructions. In comparison to Labov’s earlier analysis of Negative Inversion sentences,
Martin’s use of syntactic theory gives these sentences an entirely new structure. Martin’s
analysis argues, that the so called Negative Inversion construction in African American
English (AAE) is existential. However, unlike Labov, Martin argues that NI constructions
have only one structure and not two. Below, I will outline Martin’s two key arguments
that NI sentences are existential and not inversion constructions. Last, I will give several
hypotheses which Martin (1993) discusses in a later paper on why the expletive subject

deletes in AAE.

Martin’s argument that N I sentences are not inversion constructions is based on the

following data:

(8)
a. John ain’t believe mary liked no one (AAE)
b. * Why ain’t John believe Mary liked no one. (AAE)

c. Ain’t many doctors roun’ here (that) know (nothin’lone iota) ‘bout acupuncture.

In (8a), Martin shows that a negative polarity item (NPI) can be licensed in an
embedded clause. However, in (8b), where wh-inversion occurs, licensing of the
embedded NPI is blocked. In Ni involves inversion, then an NPI occuring in a N1
sentence should not be licensed. Nonetheless, (8c) shows that NPIs are allowed in in N I
sentences this leads Martin to conclude that NI sentences are not inversion constructions.
Below, I will lay out the details of this argument

Martin discusses the syntactic structure of simple and pleonastic negative sentences.
Simple negation has one overt or ‘inherently’ negative element such as nobody or doubt
which forms the negative sentence structure of standard English. In nonstandard English,
pleonastic negative sentences contain several negative elements within a single sentence,
where one true negative element, or “true Neg”, controls all the other negative elements
occurring in the sentence. Martin argues that the grammars of Standard English (SE) and
nonstandard English differ minimally and are interpretable to all speakers of English
because they use the same phrase marker which makes the general interpretation of negative
sentences possible (pg. 26).

Martin uses Klima’s (1966) transformational analysis to identify the “true” negative
element in pleonastic sentences. Klima’s analysis has the negative feature as the daughter

of S(entence), where it triggers lexical transformations changing what he calls the

indeterminate word class into the indeﬁnite word class. In the example below, the

indeterminate some is transformed into the indeﬁnite any (Taken from Martin 1992: 35).

(9)

a. Neg John has some money -’
Neg John has any money-*

b. Neg John ate some ketchup too-r

Neg John ate any ketchup either ->

Kihna’s rule identiﬁes the potential attachment sites of the Negative feature where
“Neg” attaches to the ﬁrst indeﬁnite in the sentence appearing before the modal or the
auxiliary. When an indeﬁnite NP does not appear before a modal or auxiliary, the negative
feature may attach to the modal, or auxiliary, then continue on to the ﬁrst indeﬁnite NP in
VP. Martin argues that the negative feature is on Inﬂ. He illustrates this by showing that
when negative polarity items (N PIs) appear in embedded clauses that are not negated, the
sentence is ungrammatical. In examples (10a - f), the matrix clause is not negated and the
embedded NPIs form ungrammatical constructions. The reason for this ungrammaticality is

due to the fact that the sentence lacks “true” Neg.

(10)

a. Patty said that Bill has (* any) money.

b. Who did Patty say has (* any) money?

c. Why did Patty say that Bill has (* any) money?

nonstandard English
d. Patty say that Bill got (* any/* no) money?
e. Who Patty say got (* any/* no) money?

f. Why Patty say Bill got (* any/* no) money?

In the second example set (1 lat), the embedded clause is afﬁrmative and the matrix

clause is negative. In this set, the embedded NPIs are possible.

(11)
a. Patty didn’t say that Bill has any money.
Nonstandard English

b. Patty ain’t say Bill got (no) money.

Martin argues that Inﬂ, where the negative feature has attached itself, is able to
license embedded NPIs because of its ability to govern the functional category that the NPI
is in. When embedded NPIs are not in the scope of the Inﬂ containing the negative feature,
the negative feature cannot be assigned. As a result, the sentence is ungrammatical because
Infl fails to license the embedded NPI. Martin argues that embedded NPIs are licensed

under the following conditions:

(12)

a. A [+Neg] Inﬂ licenses NPIs within a maximal functional category iff (if and only if) it
governs that category.

b. A functional head X° governs a functional category {YP/Y’ } iff X° m-commands YP,
and YP m-commands X°.

c. X m-commands Y iff the ﬁrst maximal projection dominating X dominates Y and there
is no ZP such that ZP dominates Y and ZP does not dominate X.

(1. m-command is relative to the type of phrase (i.e. X,Y, Z in (0) must be all lexical or all
functional).

In order for the embedded NPI to be negatively licensed by Inﬂ, Infl (the head X°)
must govern the functional category (1P) which governs the embedded NPI. In (12), (a)
and (b) requires that Inﬂ and the embedded IP have a relationship of mutual m-command.
Martin’s notion of government speciﬁes that the governing relations of functional
categories and lexical categories be treated differently. This means that intervening lexical
categories such as VP are transparent and do not affect the governing domain of functional
categories such as IP and CP. This enables functional categories to form mutually m-
command relationships with other functional categories; a condition that allows the Inﬂ
(head X° ) to license the embedded IP (YP, Y’) which assigns the negative feature to the

NPI through government as shown below.

10

(13) IP‘

/\
John I’
/\
[+Neg] Infl didn’t VP‘
/\V,
/\
believe IP2
/\
Mary I’
/\
Inf] VP2
///\\IV’
/\
liked anyone

In (13), [+neg] Inﬂ govems the IP2 because it meets the conditions of m-command.
Since VPl, being a lexical category , is invisible the maximal functional projection that
dominates IP’ is IP‘ (the matrix IP). Given the fact that IP1 dominates both Infl and 1P2,
Inﬂ and IP2 are in a relation of mutual m-command. With the conditions of m-command
being met, Inﬂ can govern into the category that the NPI is in, thus, licensing it by
assigning to it its negative feature through govemment. This predicts the grammaticality of
the sentneces in (14).

(14)
a. John didn’t believe Mary liked anyone.
b. John couldn’t stand the fact that Susan showed no pity.

c. John didn’t believe that Mary liked anyone.
Martin assumes when that appears that there is only an intermediate projection C’

and not an entire functional projection. With the complementizer phrase only projecting to

C’, Martin argues that C’ does not block Inﬂ from governing IPZ, whereas, the category

11

CP would. In other words, Martin is stating that when that appears it does not form a full

category and cannot form m-command relation with [P2 as shown in (15).

(15) 11>1
/\
John I’
/\
[+Neg] Infl didn’t VP1
Av:
. /\
beheve C’
/\
that IP2
/\
Mary I’
/\
Inﬂ VP’
/\
V’
/\
hked anyone

However, in instances where Inﬂ undergoes inversion in Wh movement

constructions, the embedded NPI is ungrammatical as shown in (16).

(16)
a. * Why didn’t John believe Mary liked anyone.
b. * Why couldn’t John stand the fact that Susan showed no pity.

c. * Why didn’t John believe that Mary liked anyone.

Martin argues that since Infl has moved from the only position where it forms a

mutually m-command relationship with the embedded IP, Inﬂ no longer governs the

embedded IP and therefore cannot license the NPI that it contains as shown in (17).

12

(17) CP

/\
Why C’
/\
didn’t 1P1
/\
John 1’
/\
t VP
AV,
/\ 2
believe IP
M /\I
ary
/\
1an VP
V,
/\

In (17), Inﬂ begins in IP1 where it can govern both the matrix clause and the
embedded clause. Movement to Inﬂ is triggered by the wh item. Subject Auxiliary
inversion occurs moving Inﬂ over the subject to C°. Martin argues that the governing
relationship between the [+Neg] Inﬂ and the embedded NPI is affected by Subject
Auxiliary Inversion (SAD. In (17), the maximal functional projection that dominates IP2 is
still IPl. However, with Inﬂ in C, the first maximal functional projection that dominates
Inﬂ is CP. Inﬂ and IP2 do not mutually m-command each other because the node that
dominates IP2 is not the same node that dominates Inﬂ, As a result, Infl does not govern
IP2 and the N PI does not receive its negative licensing.

Therefore, the examples in (18) are ruled out because there is no “true Neg” in the

embedded sentence to assign the NPI its negative feature.

13

(18)
a. Who didn’t,- [1p Patty t say [Cp has (* any) money?]] (SE)
b. Who ain’t ti [1p Patty t say [cp got (*no) money?]] (AAE)
(Martin 1992 p. 59 ex. 16)
c. Who didn’t t [11) John t ask [ [p to give (a/* any) present to Walter]]
(I. Who ain’t ti {1}) John t ask [IP (to) give {a/* no} present to Walter]]
(Martin 1992 p. 50 ex. 20)

The above licensing condition predicts that since Inﬂ has raised above the subject,
as in wh movement operations, it is not in the position to license the embedded NPI.
Now we can see how this fact can be used to show that NI constructions do not involve
inversion. In negative inversion sentences in AAE, Martin’s licensing condition predicts
that the embedded N PIs should be ungrammatical. The reason for this is simple: if
negative inversion involves movement of Inﬂ then the IP containing the embedded N PI will
no longer m-command the moved Inﬂ, just as in the wh-questions shown above.

However, NPIs are grammatical in AAE NI constructions. This is shown in (19).

(19)

a. Ain’t many doctors roun’ here (that) know (nothiti/ one iota) bout acupuncture.

b. Don’t nobody say that dealers sellin’ drug in the school yard {no more/anymore }.
(Martin 1992 p. 64 ex. 29)

For this reason, Martin concludes that NI sentences do not involve inversion.
Instead, he argues that in NI sentences, the auxiliary remains in [P1 where it can govern
both the matrix clause and the embedded clause IP2. In NI sentences, since Infl remains in

1P1 where it can govem 1P2, it can license the NPI in 1P2. Martin concludes that NI

l4

constructions are the counterparts of Standard English existential sentences which have the
expletive subject there.

Further, Martin shows that N I constructions, like existential constructions, must
adhere to the deﬁniteness restriction. This condition speciﬁes that the post auxiliary subject
must be an indefinite NP. When the post auxiliary subject is not an indeﬁnite NP, the

sentence is ungrammatical.

(20)

a. There isn’t a man in the bank. (SE)

b. Ain’t no man in the bank. (AAE)

c. Won’t nobody stay in the bank. (AAE)

d. * There isn’t the man in the bank. (SE)

e. * Ain’t the person in the bank. ( the person in not in the bank.)
f. * Won’t John stay in the bank. (John won’t stay in the bank.)

The defmiteness condition is a deﬁning characteristic of existential sentences
according to Milsark (197 2). Martin concludes from his analysis that Negative Inversion
sentences are not verb raising structures. Instead, he argues that these sentences are
negative existential sentences where the matrix Infl remains in place and retains the NPI
licensing condition described above. In nonstandard English, Martin argues that the
pleonastic subject is deleted by a PF rule which deletes the unstressed expletive subject

In a later paper, Martin (1993) explores three possibilities as to why these negative

existential sentences have a missing there expletive, and states the shortcomings of each.
His ﬁrst hypothesis assumes that there is a “null, non-theta NP” that is generated in the
Spec position of IP. There then lowers at LF to VP where it adjoins to the left of the

indefinite subject. By lowering to the subject, the null NP acts like the overt there.

15

However, Martin notes that his research suggests that there are several phonetically
null elements which have grammatical and semantic functions. Martin points out that
uncertainties arise with this analysis when trying to deﬁne a principle for ruling out the
non-negative declarative reading shown in (21a). The Null NP hypothesis predicts that,
(21a) and (21b) would have the same meaning, but in fact, (21a) can only be interpreted

as a question.

(21)
a. Is someone in the room.

b. (There) is someone in the room. (Null expletive hypothesis)

Martin’s second hypothesis assumes that the expletive there is generated in all
existential sentences and is able to delete at PF under head government.1 This analysis
assumes that the auxiliary verb raises to C and head governs the expletive. Under this
analysis, the expletive is present at both SS and LF and functions just like overt there.
Martin notes that this analysis is problematic in that no explanation can be offered to
explain why there is not able to delete in question sentences with SAI that are existential

as shown in the contrast in (22).

(22)

a. There ain’t no flowers to put on the table.

b. Ain’t there no ﬂowers to put on the table?

c. * Why ain’t e no ﬂowers to put on the table?

(1. * Ain’t e no flowers to put on the table?

 

l Head government is deﬁned as follows: Xo governs YP iff
(i)X issisterto YPor
(ii) YP is immediately dominated by Z? and X is sister to ZP.

l6

Martin’s third hypothesis involves expletive deletion with no inversion. This
analysis assumes that the movement operation of inversion does not take place and that the
expletive subject deletes.2

In this analysis, sentences with an overt there and a covert there are identical in
structure. Another problem that Martin notes is that there is no way of limiting the
deletion of there. Finally, there is nothing to explain why there is not deleted in
afﬁrmative sentences.

Unlike Labov, who states that there are two structures, the fronted negative
construction and the existential construction, Martin argues despite these problems that N I

constructions have only a single structure, namely existential.

1.3 An optimality account of NI

The most recent analysis of N I constructions is Sells, Rickford and Wasow (1996)
(hereafter Sells et al.). Like Labov, Sells et a1. argue that NI constructions must have two
analyses to account for sentences having either the copula-be auxiliary or a modal
auxiliary. Sells et a1. argue, following Labov, that NI sentences with the copula-be
auxiliary are existential. However, they conclude that NI sentences are not inversion
constructions but are non-movement constructions. Sells et al.’s analysis proposes that the
following structures (23) and (24) correspond to Labov’s existential sentences and

inversion sentences respectively.

 

In English. sentences are required to have an overt subject rn every clause. Tlus requirement rs called the Extended PIOJCCUOII
Principle.

17

(23)
IP (inversion)

NP 1’
can’t VP

[NP/\V

nobody tag you then

(24)
IP (existential)

NP 1’

Arn’t NP
/\
NP C
nothin’
I

(you can do) (Sells et a1. 1996 p. 606)

Instead of an inversion construction, Sells et a1. conclude that the subject remains
internal to VP and does not raise to the clausal subject position, [Spec, IP], as shown
above in (23). Here Labov’s inversion construction structure is replaced by the internal
subject analysis.

Sells et al. show that NI sentences with modal auxiliaries do not raise from I to C.
They illustrate this by showing that the complementizer, that, and the negative auxiliary

appear simultaneously in embedded clauses.

18

(25)
a. Bill said that ain’t nobody in the room.

b. Bill said that can’t nobody leave.

Under inversion, we expect the modal auxiliary to move to C. However, when
that appears in an embedded clause, C is ﬁlled. With C being ﬁlled by that, the auxiliary,
which is to the right of that, has not risen to C. Thus, inversion is ruled out. As a result,
Sells et a]. argue that neither the existential nor the “NI construction” involves movement.
In addition, Sells et 31. state that neither the existential structure nor the inversion structure
has anything generated in the canonical subject position. To account for this fact, a theory
of optimality places constraints on the expression of negation in addition to other
constraints which allow NI expressions to exist (Sells et al., 1996:596).

In Optimality Theory, constraints are ranked, which allows the lowest ranking of
two constraints to be violated if the higher ranking constraint is satisﬁed. Sells et al.

propose the following two constraints to account for NI constructions.

(26)
(1) Negative quantiﬁers in AAVE must be c-commanded by a negative auxiliary,

(11) Spec IP must be ﬁlled.

Ranking constraint (I) over constraint (11), allows the canonical subject position [Spec, IP]
to be unfilled thus accounting for both the Internal Subject and existential analysis .

The ﬁrst constraint, (1), requires that the post auxiliary subject be c-commanded by
a negative auxiliary. The condition of c-command restricts the post auxiliary subject from
moving to the left of the negative auxiliary. In (27 b and c), the sentences are
ungrammatical because the subject c-commands the negative auxiliary, and violates

constraint (1).

19

(27)
a. Can’t nobody afford these high prices. (AAE)
b. * Nobody can’t afford these high prices.

c . * Nobody ain’t in the room.

The second constraint, constraint (1]), requires that the canonical subject position,
[Spec, IP] be ﬁlled. This constraint will force Spec IP to be ﬁlled in afﬁrmative sentences
such as (28a and b) and rules out (280 and (1) because constraint (1) is not applicable, and

constraint (ii) therefore cannot be violated.

(28)

a. There’s a man in the room.
b. Somebody can leave now.
c. * Is a man in the room.

(1. * Can somebody leave now.

For example, in sentence (29) below, Sells’ et al.’s analysis argues that the
negative auxiliary is in the head of IP which meets the conditions of constraint (I) but
violates (II). This sentence is grammatical in that the primary constraint, (1) has been met.
Though the Spec position of IP is unﬁlled, the post auxiliary subject is c-command by the
negative auxiliary positioned in [IP]. In (30), the same holds true. The post auxiliary

subject is c-commanded by the negative auxiliary in IP.

20

(29) [IP Spec [can’t [VP nobody tag you then]]]

(30) [IP (there) [ain’t [VP nothin’ went down]]]

Sells et a1. provides further evidence to show that certain negative inversion
constructions cannot take on expletives and thus are analyzed as non-existential. Examples
(31a and b) are ungrammatical when the expletive is added as (31c, (I) shows. On the other
hand, in (32), the sentences remain grammatical with the expletive present Sells et al. give

these sentences an existential analysis.

(31)

a. Can’t nobody tag you then
b. Don’t nobody see it.

c. * It can’t nobody tag you.

(1. * It don’t nobody see it

(32)

a. (There) Ain’t nothing’ happen’

b. (There) ain’t nobody goin’ out town.
c. (It) ain’t nothin’ worth cryin’ about.

d. (There) ain’t no stop sign here.

Although Sells et al. have established that NI sentences are not inversion
constructions, their hypothesis maintains that there are two analyses of NI. For both
analyses, the negative auxiliary is in the head of IP. However, in the non-existential
construction, the post auxiliary subject is in VP, while in the existential constructions, the

post auxiliary subject is an NP with a reduced relative clause.

21

(33)

a. Ain’t nothin’ went down.

b. There ain’t nothin’ that went down.

c. Ain’t nobody ever thought ‘bout picking up nothin’.

(1. There ain’t nobody that ever thought ‘bout picking up nothin’.

The two analyses differ according to whether the sentence is a monoclausal
structure (where the predicate is the main clause), or whether the predicate is in a reduced
relative clause, a biclausal structure. Sells et al. state that existential constructions are
biclausal structure and that inversion constructions are monoclausal. Sells et al. along with
Labov et al. (1968), argue that sentences such as (33) are biclausal constructions. In (33a),
nothin’ went down is treated as a relative clause where the relative pronoun that has
been deleted. Sells et al. argue that the existential analysis contains two separate clauses
where the relative pronoun is deleted. Further, in (34), Sells et a1. gives examples from
Guy and Bayley (1995) which show that deletion of relative pronouns does occur in other

instances in AAE.

(34)
a. Miss Rushkin the one o help me get into this program.
b. What’s the worst thinge can happen?

c. I don’t know what the old woman’s name e that done the, the cooking.

(Sells et al. 1996, p.602)

Also, there are other examples which show that the post auxiliary is not the
underlying subject (Sells et al. 1996, p. 603). In these sentences, the underlying subject

appears to be in another clause.

22

(35)
a. Ain’t nothin’ you can do for ‘em.

b. Ain’t no way in the world you can miss it.

In conclusion, Sells et al. state that neither the inversion nor the existential sentence
structure involves Auxiliary movement Their analysis states that NI sentences with
negative existential verbs are existential constructions with null expletives. These sentences
are biclausal. Their second analysis states that non-existential sentences are formed with

negative modal auxiliaries. This structure allows the subject to remain internal to VP.

1.4 Weldon’s account of NI sentences

Weldon (1996) argues that two analyses are needed to explain Negative Inversion
sentence structures in African American English as well. She holds that in order to account
for NI sentences, the inversion and the existential analyses originally proposed by Labov
must be maintained. Second, Weldon convincingly argues that NI constructions do not
contain a reduced relative clause as Sells et al.’s (1996) existential analyses proposes. By
demonstrating that NI sentences are monoclausal and not biclausal, she concludes that the
relative pronoun that is not deleted. Furthermore, she shows that the relative pronoun
never appears in N I sentences. For these reasons, Weldon argues that NI sentences are
inversion constructions. She also shows that covert expletives are ungrammatical in certain
NI constructions containing copula be. In these cases, Weldon builds onto her argument

that an inversion analysis should be maintained.

1.4.1 Existential Quantiﬁers
In NI sentences, Weldon argues that the post auxiliary subject is not affected by
negative attraction. Thus, in the event where an indeﬁnite like a man is transformed into a

negative polarity item such as no man, she holds that the subject remains positive. She

23

argues that under the Head-Driven Phrase Structure Grammar Theory (HPSG)3, the
auxiliary selects an NP which only appears to be a negative existential quantiﬁer (NPI).

She argues that it is due to negative attraction and/or the fact the subject is c-commanded
by the negative auxiliary, that the subject appears to be an NPI. From this, she holds that
NI’s post auxiliary subject maintains its positive attributes. Weldon’s re-analysis of the
post auxiliary subject is based on a phenomenon occurring between the negative quantiﬁer
few and the positive quantiﬁer many. She predicts that negative concord (or negative
attraction) changes a positive indeﬁnite into the N PI few making the post auxiliary subject
a negative quantiﬁer. Weldon’s conclusion suggests that the negative concord rule rejects
the negative quantiﬁer, few, for its positive counterpart many. As a result she argues that

the negative quantiﬁer in NI sentences are really a positive existential quantiﬁer.

(36) * Don’t few people live around here.

(37) Don’t many people live around here.

Weldon’s argument that negative quantiﬁers remain positive is based on the
ungrammaticality that occurs with few. In (36), the quantiﬁer “few people,” does not
appear grammatically in NI constructions while the positive quantiﬁer auxiliary, “many
people,” does. Weldon predicts that the existential quantifier, few, being negative, should
behave as other negative existential quantiﬁers do in NT construction. However, this not
being the case, Weldon reasons that since its positive counterpart many is grammatical in

NI constructions that all existential quantiﬁers must be positive.

1.4.2 The Non-copular Case
Weldon states, following Labov (1972), that expletives are optionally null only in

the copular construction and obligatory in non-copular constructions. Here, she notes that

 

3 The HPSG is a non transformational. The details of Weldon's analysis are not relavant for the discussion
here. See Pollard, and Sag (1994).

24

Sells et al.’s and Martin’s models do not offer an explanation as to why their analyses
allow the formulation of ungrammatical sentences involving non-copular verbs (38 is one
construction that Martin’s analysis does not account for). Based on the inadequacies of

their arguments, Weldon argues that an inversion analysis must be maintained.

(38) * It/there don’t nobody like me.

Instead of a null expletive analysis, she proposes that N I sentences are related to

non-inverted sentences such as in (39).

(39) Nobody don’t like me.

1.4.3 The Copular be; Verb

Weldon distinguishes four types of copula be (the negative auxiliary ain’t). Of
the four, there is only one that is analyzed as being existential.
(40)
Present Copular ain’tl It/There ain’t no black Santa Claus.

ain’t; Ain’t nobody a man.
Present Perfect ain’t3 Ain’t nobody been to the store yet.

9

Past Tense ain t4 Ain’t nobody see it coming.
do-support

Of the copular auxiliaries, Weldon has categorized the copula be into four types of
ain’t which are separated into three groups by tense. Of the three tenses, two of the four
types of ain’t are present tense. The ﬁrst copula bel (41a) selects an optional expletive
subject. The be; type (41b) selects a non-existential subject. The copula beg type (41c) is

used in conjunction with a present perfect verb. The be4 type (41d) is the past tense

25

auxiliary used instead of do in AAE. Weldon argues that be] appears in existential

constructions and be2 in NI constructions.

(41)

a. (It/There) Ain’tl no back Santa Claus.

b. (?It/* There) Ain’t; nobody a man.

c. (* It/There) Ain’t;, nobody been to the store yet.
(:1. (* It/there) Aint’t4 nobody see it coming.

In building her case for inversion (see below), Weldon notes that (42), under Sells
et al.’s (1996) existential analysis, can be either analyzed as an inversion or an existential
construction. Sells et al’s analysis would predict that (42) would appear as (43), a
biclausal construction. One would then have to assume that lexical items in CP (in this case
who) are deleted. According to Weldon’s hypothesis, the auxiliary is of the ain’t; type
which selects a non-existential subject Therefore, giving it an inversion analysis , (42)
would pattern as (44), a simple clause construction. As an inversion construct, this avoids
the assumption that who has been deleted, and the problem of hypothesizing what

motivates its deletion.

(42) Ain’t nobody ever thought about picking nothing up. (NYC)
(43) (It/There) ain’t nobody (who) ever thought about picking up nothing.
(44) Nobody ain’t ever thought about picking up nothing.

However, Weldon shows that (42) is not a biclausal construction. As shown in
(45 a, b), wh-extraction possible out of monoclausal constructions but not out of reduced
relative clauses. The grammaticality of (450) illustrates that NI constructions allow wh-

extraction. This proves that N I constructions must be monoclausal. However, under Sells

26

et al.’s biclausal construction, NP extraction should be prohibited. Since under Sells et
al.’s analysis the reduced relative clause functions like a complex NP, and extraction out of
a complex NP is not possible, this demonstrates that NI sentences are monoclausal or

“simple clauses”.

(45)
a. This is the thing t,- [s that nobody ever thought about picking t, up].
b. * This is the thing [3 that there ain’t nobody (who ever thought about picking ti up].
c. This is the thing, [3 that ain’t nobody thought about picking ti up].
(Weldon 1996 p. 12 ex. 69-72.)

Weldon presents another argument for the position of inversion. Milsark (1977),
showed that existential constructions disallow non-state-descriptive (individual level)
predicates. Since individual level subjects are ungrammatical in existential sentences and
grammatical in NI constructions, she reasons that NI sentence must be inversion

constructions.

(46)
a. (There) Ain’t nobody available (stage level)
b. (* There) ain’t nobody intelligent. (individual level)

c. (* There) ain’t nobody a man. (individual level)

(47)
a. Nobody ain’t available.
b. Nobody ain’t intelligent.

c. Nobody ain’t a man.

27

The examples above show that when the expletive is added to the NI construction
containing an individual level predicate as the post auxiliary NP, the sentence is
ungrammatical. Weldon contends that (46 b and c) are inversion constructions which relate

to the non-inverted sentences of (47 b and c) respectively“.

1.4.4 Ambiguous Cases

Weldon’s last argument suggests that there are some cases where determining the
structure of certain sentences is difﬁcult. Weldon claims that sentences such as (48) can be
interpreted as an inversion construction or as an existential construction. In (48), Weldon
hypothesizes that the copula he could either be the ain’t; type or the ain’tl type. When
the sentence has the ain’t; type, she states that the auxiliary chooses a nominative subject,

thus designating an inversion analysis. On the other hand, with the ain’t, type, the

auxiliary selects an optional existential subject, and the existential analysis applies.

(48)
a. Ain’t nothin’ broken on me to ﬁx.
b. I knew wasn’t no Melinda in here.

c. Wasn’t nobody gettin’ hurt or nothin’.

(Weldon 1997 p.14)

Given the fact that covert expletives are grammatical in these constructions, Weldon
suggests that these sentences can either express an inversion construction, or existential
construction. The inversion construction has a declarative reading which depicts a state that
is deﬁnite or a state that is exact such as John is in the room, opposed to, John is not in the
room. Existential sentences, on the other hand, express that of all the people who are in the

room there is no person named John in the room: There isn ’t a John in the room.

 

4 Though Weldon predicts that ($46b and c) are the non inverted counterpart of ($47b and c), I ﬁnd ($47b
and c) to be ungrammatical.

28

In sentences where the interpretation can either have an existential analysis or an
inversion analysis, Weldon argues that these sentences can be disambiguated by
determining the structure of the grammatical category modifying the post auxiliary subject.

She attempts to illustrate tln's in the following manner.

(49)

a. Ain't nobody that's happy rich.

b. Nobody that's happy ain't rich.

c. It/There ain't nobody that's happy rich.

In (49b), Weldon states that the interpretation would read as Nobody that's happy
is rich, as an inversion construction. As an existential construction, the interpretation

would read, There isn't anybody who is happy being rich. for (49c).

1.4.5 Existential Cases

Existential cases involve the auxiliary be] and are followed by a single NP
complement. Weldon states that in existential sentences, the expletive is optionally null and
utilizes Martin’s unstressed analysis as a possible theory to explain why the expletive
deletes. Last, Weldon argues that the subjects in these sentences must be a quantiﬁed NP.

These expletive subjects are only optional with negative auxiliaries.

29

(50)

a. Aint no black Santa Clause.

b. Wasn’t no such thing as :’Well, I didn’t do it.’
c. Wasn’t nobody I could trust

d. Ain’t nothing you can do for them. (Weldon 1997 p.13)

Under an inversion analysis, the above examples are ungrammatical. Weldon

holds that the ain't1 chooses an existential subject and that sentences with the ain’tl are

ungrammatical inn the non-inverted sentence construction. Weldon's argument indicates

that since the ain'tl type verb chooses an optional existential subject, (the expletives there

and it,) these subjects cannot appear as the subjects of inverted sentences.

(51)

a. * N 0 black Santa Claus ain’t.

b. * No such thing as : ‘Well, I didn’t do it.’
c. * Nobody I could trust wasn’t

d. * Nothing you can do for them ain’t .

(Weldon 1997 p. 13)

In other words, when ain'tl type verbs select an optional existential subject, an
existential sentence construction is formed, and use of the expletive subject is optional.

In order to account for inversion in embedded clauses, Weldon assumes
Culicover’s (1991) analysis of negative inversion in Standard English (SE). In SE,

sentence such as (52) where the negative has been fronted are grammatical.

30

(52)
a. I said that under no circumstances will I talk to Bill.

b. I said that never again will I talk to Bill.

Culicover argues that there is an extra position existing between CP and IP, which he calls

Polarity Phrase (PolP) which is the landing site of the fronted negative as shown in (53).

(53) CP
/\C’

/\
that POIP
/\

(Neg) never again PolP
/\
I IP
willi /\
P I’
’I /\
ti VP

to Bill

In (53), the auxiliary raises above the subject to head PolP where the fronted negative
resides (in [Spec, PolP]).

1 .5 Summary
In retrospect, we have learned that there are several points that hold consistent in the
literature on NI constructions in AAE thus far. One consistent aspect of the literature has

been the method of determining whether a particular sentence is existential or inversion: In

sentences where the expletive can appear, the sentence construction is given an existential

31

analysis. On the other hand, in sentences where the expletive cannot appear, the sentence is
given an inversions analysis.

Also, we have learned that in NI constructions, the post auxiliary subject must
adhere to the deﬁniteness effect. This condition requires the post auxiliary subject to be
indeﬁnite. Further, it has been shown that the negative auxiliary does not move to CP.
This is illustrated in embedded sentences where the negative auxiliary appears along side
the complementizer that. However, we found through Weldon, following Culicover
(1991), that there is an additional node existing between IP and CP following her inversion
analysis. As a result, the landing site for the auxiliary could be PolP, instead of C. PolP
being a polarity Phrase, would license the negative feature, thus, the auxiliary raises to
PolP in order to form NI expressions

Last, it has been convincingly argued that N I constructions are not biclausal
constructions with reduced relative clauses. This rids us of the misnomer that NI
constructions delete the subject complementizer that from CP, in hopes of maintaining an

existential account of NI sentences.

32

CHAPTER 2

The aim of this chapter is to outline the major problems that exist in the current
analyses of NI constructions. In the previous literature, there are several arguments which
hold that there is a need for two structures to account for NI constructions. I will argue
that NI constructions have only one structural analysis, this structure being existential. The
chapter begins with the discussion of Labov's analysis, which contends that NI sentences
are like fronted negative sentences in Standard English. I will show that fronted negative
sentences differ from N1 sentences in respect to the indefmiteness restriction.

Next, I will discuss the shortcomings of Sells et al.'s Internal Subject analysis.
Here, problems arise when considering c-command. Sells et al.’s analysis allows the
ungrammatical construction of sentences which c-command does not rule out. Sells et al.’s
analysis also does not consider the signiﬁcance of deﬁniteness restriction.

Areas of Martin’s and Weldons’s analysis, when tested, show that several aspects
are problematic. In the following, I will show that Martin’s licensing condition does not
rule out NPI’s in embedded clauses. Also, I will illustrate that Weldon’s analysis is
problematic for the simple fact that she does not recognize that NI sentence must adhere to

defmiteness restriction.

2.1 Problems with N I sentences as Fronted Negative constructions

Labov’s analysis states that sentence with the verb ain’t appearing initially are
existential sentences with null expletives. The other sentence type has a modal verb
appearing initially and is said to pattern after SE’s fronted negative sentences where subject
auxiliary inversion (SAD occurs. For example, in the sentences below, the auxiliary
moves above the subject in front of the negative adverbial. Fronted negative sentences are
derived from non-inverted sentences. In the examples below, (54a) is the derivative of

(54b).

33

(54)
a. Never again will Susan trust another man with her heart.

b. Susan will never trust another man with her heart again.

In Labov’s analysis, there is no formal syntactic comparison made between NI
constructions and fronted negative sentences in Standard English. He assumes that both
the NI constructions and the fronted negative constructions have the same movement
operations of SAI. If AAE N I sentences are fronted negative sentences, we would expect
the verb to move from I° to C’. However, there are several arguments which illustrates
that an inversion analysis is highly unlikely. First, Sells et al (1996 p.603) show that if NI
sentences involve movement from 1’ to C’, they can not occur in embedded sentences.
Under the inversion analysis, this movement would be blocked when the complementizer
that is present in C. The complementizer, that when present would occupy the only

position where the auxiliary could land.

(55)

a. I know a way that [can’t nobody start a ﬁght].

b. Pilate they remembered as a pretty woods-wild girl “that [couldn’t nobody put shoes
on].” (Sells et al., p.603)

However, N 1 sentences do appear grammatically in embedded sentences containing the
complimentizer that, therefore, the auxiliary in NI sentences does not undergo 1° to C
movement.

There is further evidence to suggest that the auxiliary in NI sentences do not

undergo inversion. The question phrase how come does not allow I0 to Co subject-aux

34

inversion. This is shown in (56c and e). On the other hand, NI sentences are acceptable

with how come as (56a, b and (1) show.

(56)

a. How come ain’t no ﬂowers on the table.

b. How come there ain’t no ﬂower on the table.
c. * How come ain’t there no ﬂower on the table.
d. How come John didn’t leave.

e. * How come didn’t John leave.

Second, if we assume that NI constructions are fronted negative constructions, we
should expect sentences with non inverted auxiliaries to be grammatical, as non fronted
negative sentences are in SE. In the examples below, we see that fronted negative
sentences do not behave as NI constructions do. This suggests that AAE, NI sentences

and Fronted Negative Sentences are structurally different.

(57)

a. Can’t nobody go to the store.

b. Can’t no boy wear a dress and not be noticed.
c. * Nobody can’t go to the store.

(1. * No boy can’t wear a dress and not be noticed.

In example (57), the negative quantiﬁer cannot form a grammatical sentence when it
is to the left of the negative auxiliary. In fronted negative sentences, the movement of the
negative is stylistic. If NI sentences are similar, then any grammatical fronted negative
sentence should be a grammatical N I sentence. The examples in (5 8) are acceptable fronted

negative sentences in SE, along with their non-fronted equivalents.

35

(58)

a. Never has Utah beaten the Bulls in the ﬁnals.

b. Utah can’t beat the Bulls in the ﬁnals.

(1. Not once has John forgotten to say his Sunday prayer.

e. John has not once forgotten to say his Sunday prayer.

In SE, fronted negative sentences are also quite restricted. For example, the simple
negation not cannot front, so the fronted version of (59a, and c) are not possible, as

shown in (59 b, and d).

(59)

a. Utah can’t beat the Bulls.

b. * Not can Utah beat the Bulls.

c. John can’t forget to say his prayers

d. * Not can John forget to say his prayers.

In addition, AAE NI sentences are unacceptable in SE (e. g. 57 above).
However, in NI constructions, the negative auxiliary cannot raise above certain
types of subjects. In a sentence with a deﬁnite NP as its subject (60), the sentence is

un grammatical.

36

(60)

a. * Can’t Utah beat the Bulls in the ﬁnals.
(meaning Utah can’t beat the Bulls in the ﬁnals.)

b. * Can’t John forget to say his Sunday prayer.
(meaning: John can’t forget to say his Sunday prayer.)

c. * Can’t John never forget to say his Sunday prayer.
(meaning: John can never forget to say his Sunday prayer.)

In the above examples, we see that N I sentences and fronted negative sentences
behave differently in terms of inverting over deﬁnite NP subjects. In (60a) the non-fronted
sentence reads, “Utah can’t beat the Bulls in the ﬁnals.” If we compare the sentence
structures of (59a and c) to (60a, b, and c) we will notice that NI constructions have a strict
adherence to the deﬁniteness effect The subject NPs that are ungrammatical in the
sentences above, are all deﬁnite NPs. Milsark (1977) shows deﬁnite N Ps have the same

ungrammatical effect in existential sentences as shown in (61).

(61)

a. There is a wolf at the door.

b. * There is the wolf at the door.

c. There were several people cycling along the creek.

d. * There were John and Mary cycling along the creek
e. There was an article mentioned.

f. * There was Frank’s article mentioned.

(Milsark 1977, p. 4. ex 5a-f)

It is interesting to note that (60a), for example, can turn into an. acceptable NI

construction if the subject is negative, as is in (62).

37

(62) Can’t no Utah beat the Bulls in the ﬁnals.

However, negating the subject also makdes it indeﬁnite with the meaning "any Utah
team", thus providing further evidence for the deﬁniteness restriction.

Even in NI constructions that do not contain the verb be, the conditions of
deﬁniteness restriction must still be met. The examples below demonstrate that a N1
sentence must meet the conditions of the deﬁniteness restriction. The fact that fronted
negative sentences lack this condition, as the contrast between (58) and (60) shows,
demonstrates that NI sentences are not fronted negative constructions. In AAE NI
constructions, the NP must be indeﬁnite or the result is ungrammatical as shown in the

contrast between (63) and (64).

(63)

a. Can’t nobody stand to be in this rut again
b. Can’t nobody land on the moon.

c. Can’t nothing you do hurt me.

(I. Can’t nobody root for the Boston Celtics.

(64)
a. * Can’t I stand to be in this rut again.
b. * Can’t the Russians land on the moon.

* Can’t I root for the Boston Celtics.

0

P-

* Can’t you hurt me.

The fact that the defmiteness restriction is required in NI sentences makes Labov’s
inversion analysis questionable. If NI sentences are inversion constructions, why must

defmiteness restriction be a factor in determining its grammaticality?

38

2.2 Problems with Sells et al. 1996

Sells et al. (1996:596) also reject Labov’s inversion analysis and they argue that
the negative quantiﬁer must be c-commanded by the negative auxiliary to account for NI
sentences structure. In dealing with sentences containing non—copula verbs, this condition
acts as a constraint which enables the subject to remain in VP. However, the
ungrammaticality of (65) is problematic for Sells et al.’s analysis. Under their analysis,
(65) should be grammatical; the negative quantiﬁer is c-commanded by the negative
auxiliary meeting constraint (1). However, the sentence is still ungrammatical due to the
deﬁniteness restriction. Like existential sentences, the deﬁnitenes restriction only shows
up on the post-auxiliary subject. This illustrates that NI sentences behave more like

existential sentences than fronted negative sentences.

(65) * Can’t the bulls beat no team from Mars

IP
I?
./\
I VP
Can’t
NP V’
the Bulls /\
V NP

beat
no team from Mars

Sells et al.’s analysis predicts that NI sentences such as these would be
grammatical with deﬁnite nouns as the post auxiliary subject because the negative auxiliary
does c-command a negative quantiﬁer, the NP “no team from Mars”. The tree in (65)

shows that the conditions of c-command are being met. The branching node that dominates

39

negative auxiliary is IP', and this dominates the negative quantiﬁer no team from Mars.
Their analysis has no way of accounting for the deﬁniteness restriction that is on the post
auxiliary subject

Sells’ et al. internal subject analysis argues that NI constructions with non-copula
verbs do not involve movement of the subject. This condition of non-movement relies on
the ability of the negative auxiliary to c-command the negative quantiﬁer which restricts
the subject from moving out of VP. As a result, the negative quantifier can never appear to
the left of the negative auxiliary, the subject position [Spec, IP]. The ungrammaticality of
(65) cannot be accounted for by c-command. What rules out (65) is that the definiteness
effect is not satisﬁed.

Returning to Sells et al.’s existential analysis, there is another area where
problems occur. Sells et al.’s analysis holds that NI sentences which fall under the
existential analysis are biclausal constructions with a reduced relative clause as shown in

(64).

(66) There ain’t nothin’ that be happenin’.

IP
/\
There I’
ain’t NP
/\
N CP
. /\
n thrn’ (that) IP
(be) VP
happenin’

However, as the discussion of Weldon (1996) in chapter 1 has shown, the reduced
relative clause analysis cannot be correct. To sum up, Sells et al.’s analysis has serious

problems and thus cannot be adopted.

40

2.3. Problems with Martin’s Definiteness Effect and Null Expletives

Martin argues that because NPIs are licensed in embedded clauses in NI sentences
such as (67), NT cannot involve auxiliary inversion (AI). In Wh-constructions, AI
hinders Inﬂ from assigning its negative feature to the embedded NPI. In NI sentences, the
N PI appears grammaticality in the downstairs predicate. Because Inﬂ is able to assign its

negative feature to the NPI, this shows that N I sentences do not involve AI.

(67)

a. Ain’t many doctors around here (that) know {nothin/one iota} bout (no)acupuncture.
b. Don’t nobody say that dealers sellin drugs in the school yard {no more/any more}

c. Can’t a man in this place say he happy bout {nothin/a damn thing}

(I. There aren’t many doctors around here that know anything about acupuncture.

e. There isn’t anybody that says dealers are selling drugs in the school yard anymore.
f. There isn’t a man in this place who can say he’s happy about anything.

(Martin 1996 p. 64 ex. 29)

The main problem with this analysis is that Martin presumes that NI sentences are
biclausal constructions. Martin’s entire argument relies on this premise in fact.

Contrary to what Martin claims, and as Weldon proves with the NP extraction
test, NT sentences are monoclausal constructions and not biclausal constructions.
Therefore, the licensing conditions that Martin applies to embedded NPIs cannot apply to
N I sentences where there is only one clause. As long as the NPI is within the same clause
that Inﬂ is in, Martin’s licensing condition allows Inﬂ to assign its negative feature.
Therefore, Martin’s licensing condition cannot rule out inversion as a possible analysis.

In addition to the mechanical problems that Martin’s analysis has, it fails (as
Weldon also notes) to account for NI sentences containing initial non-copular verbs.

Though he points out that the deﬁniteness effect applies in NI constructions, he fails to give

41

an account of how these sentences can be existential constructions when they form

ungrammatical constructions when the expletive is added.

(68)

a. Can’t nobody see us talking.

b. Can’t nobody even create the mess you got us in.
c. * It/T here Can’t nobody see us talking.

d. * Itlthere can’t nobody even create the mess you got us in.
2.4 Problems with Weldon’s HPSG Account

Weldon’s (1993) HPSG analysis of AAE NI constructions (along with Labov)
concludes that there must be both an existential and an inversion analysis to account for NI
sentences. Weldon argues that sentences with non-copular verbs, and one particular case
involving ain’t are inversion constructions. She argues that sentences with overt
expletives are ungrammatical because they derive from non-inverted constructions.
However, (like Labov) Weldon does not address why the subject in NI must be an
indeﬁnite NP. Sentences with deﬁnite N Ps pose the same problem for her analysis as they
do for Labov’s. In all NI sentences containing both copular and non-copular auxiliaries,

the result is always ungrammatical when the subject is deﬁnite.
(69)
a. * Ain’t John in the room

b. * Don’t the military kill innocent people.

Furthermore, her analysis does not account for non-inverted ungrammatical

sentences involving deﬁnite NPs, as in (70).

42

(70)
a. * No team from Mars can’t beat no Bulls.

b. * Nobody can’t remain in the room.

Weldon tries to annul the deﬁniteness effect by re analyzing the post auxiliary
subject as a positive quantiﬁer. She reasons that the post auxiliary appears negative because
it is governed by a negative auxiliary. The evidence that she provides involves the positive
quantiﬁer,’ many people’, in N I sentences. Since the negative counterpart, ‘few people’,

is ungrammatical in NI constructions, she argues that post auxiliary NPs are not negative.

(71)
a. Don’t many of them live around here.

b. * Don’t few pe0ple live around here.

Weldon’s argument is based on the relationship between any and no. In SE,
when the auxiliary is negative the quantiﬁer that the auxiliary chooses is the indeterminate
any. In AAE, the opposite occurs; when the auxiliary is negative, the auxiliary chooses

the negative indeﬁnite no.

(72)
a. John didn’t do anything. (SE)
b. John didn’t do nothing (AAE)

With the quantiﬁers many and few, Weldon expects the same relationship to
hold true. With the negative auxiliary, she predicts that AAE will choose the negative item
few. However, this is not the case. In AAE, the relationship that the quantifiers many

has with few does not pattern after the relationship that no has with any. The examples in

43

(73) show that many and few function the same in AAE as they do in SE. In AAE, as in
SE, the negative quantiﬁer, few, illustrated in (7 3c), is ungrammatical within a negative
auxiliary construction. Only the positive quantiﬁer forms a grammatical sentence with the

negative auxiliary, as (73d) illustrates.

(73)

a. Many people will die. (SE)

b. Few people will live. (SE)

c. * Few people won’t live. (AAE)
(1. Many people won’t live. (AAE)

The difference in the relationship that any and no have with few and many is
shown in (74). The quantiﬁer many and few can appear as subjects. On the other hand,
while no (74d) can be the subject of a negative sentence, the indeterminate any, in (74c),
cannot form a negative sentence. However, any can appear as the subject in positive
sentences containing modals (74c). The contrast between (74b and c) shows that few

does not behave like any.

(74)

a. Many people left.
b. Few people left.
c. * Anybody left.
(1. Nobody left.

e. Anyone can leave.

Furthermore, the negative quantiﬁer, few people, cannot form a grammatical

negative existential sentence in either SE or AAE. ‘Few people’ is a problem for negative

existential sentences in general, and affords Weldon a poor position to base her argument

on.

(75)
a. There aren’t many doctors that live around here.

b. * There aren’t few doctors that live around here.

There are some existential sentences in AAE were the negative auxiliary and the

negative quantiﬁer few do appear in the same sentence.

(76)
a. * It ain’t a few women in here
b. (It) ain’t but a few women in here.

c. * It ain’t but many women in here.

However, (76a and b) illustrate that few cannot appear grammatically without the
conjunction, but. This indicates that few may not be within the scope of the negative
auxiliary. Weldon’s basic claim is that in sentence such as (77) where the expletive cannot

occur, an inversion analysis must be used.

(77)

a. Can’t nobody tag you then

b. Before they invented them kicks, couldn’t nobody do that.
c. Won’t nobody catch us.

(1. Don’t nobody break up a ﬁght.

e. Didn’t nobody see it; didn’t nobody hear it.

45

In sentences where the expletive would yield an ungrammatical sentence involving
copular verbs, Weldon argues that these are also inversion constructions. She holds that
these sentences contain predicate NPs which cannot appear in existential sentences.
Weldon notes that Milsark (1977) discovered that existential sentences only allow simple
NPs in the post auxiliary position. Weldon provides examples of NI constructions

containing the verb ain’t, where ungrammaticality occurs when the expletive is added.

(78)

a. Ain’t nobody ever thought about picking up nothing.

b. Ain’t no cop never beat me in my head.

c. Ain’t nobody never told me what to do.

(79)

a. * There/it ain’t nobody ever thought about picking up nothing.
b. * There/it ain’t no cop never beat me in my head.

c. * There/it ain’t nobody never told me what to do.

In order for the expletive to appear grammatically in these sentences, the analysis
which views NI sentences as biclausal construction (which Weldon convincingly

extinguishes) has to be applied.

(80)
a. There/It ain’t nobody who ever thought about picking up nothing.
b. There/It ain’t no cop who ever beat me in my head.

c. There/It ain’t nobody who ever told me what to do.

46

Weldon’s also argues that an inversion analysis applies to certain sentences

containing the copular-be. This particular sentence contains the ain’tz type verb which

selects a individual level predicate.

(81)
a. Ain’t nobody a man.

b. Ain’t nobody gon’ let you walk all around town to ﬁnd somebody to whip them.

Following Milsark (1977), Weldon argues that existential constructions disallow

individual level predicates. As a result, the following are ruled ungrammatical.

(82)
a. * There/It ain’t nobody a man.
b. * There/it ain’t nobody gon’ let you walk around town to find somebody to whip

them.

However, it is questionable whether or not these sentences actually yield
ungrammatical sentences. As a speaker of AAE, two of the three examples that Weldon
provides are grammatical with the expletive (the same judgments were given by other
speakers of AAE). With some examples, the expletive it is preferred over there. In
(82d), the sentence containing it sounds worse than the sentence containing there. This
sheds new light on Martin’s analysis of sentences containing there as opposed to it. In
some sentences, as (83a and b) illustrate, it and there are used synonymously, while in

other sentence constructions they contrast entirely as in (83c and d).

47

(83)

a. It is a man in the room.

b. There is a man in the room
Context : Do you have any scissors?
c. There is a pair on my desk.

(1. * It is apair on my desk

Though, arguably, some of these constructions do not accept overt expletives,
Weldon does not address why the subjects in each sentence must be indeﬁnite. If these are
truly inversion constructions, the definiteness effect should have no consequence.
Nevertheless, evidence demonstrates that this is not the case. In (83), the result is

ungrammatical with deﬁnite post auxiliary subjects.

(84)

a. * Ain’t John sittin’ in the back seat. (meaning John is not sitting in the back seat)

b. * Ain’t the Wilsons going to the show with the Hobbs. (meaning; The Wilsons are not
going to the show with the Hobbs)

Weldon’s analysis , following Culicover (1992), also seeks to remedy the IP to CF
movement operation by suggesting that is another maximal projection, the polarity phrase

(PolP), appears between IP and CP.

(85)
a. Which books did Lee say [Cpthat [90“, only to Robin ]will [IpShC give]]].
b. These are the books [prhich [Pulp only to Robin] will [1p Lee give]]].

c. Lee wonders [prhy [MP in no way] woulth Robin volunteer]]].

48

Culicover shows the negative inversion can occurs within wh—construction.
Culicover notes that the moved or fronted negative appears to the right of the
complementizer that and before the raised IP. Culicover argues that this shows at C can
take as it complement a maximal projection, PolP, which is distinct from IP. This
projection can have as its head Neg , which agrees with a negative in [Spec, PolP].
Though Culicover’s convincingly show that PolP exist as a possible landing site for
fronted negative constructions, Weldon does not show why the subjects of AAE NI
sentences require their subjects to be indeﬁnite quantiﬁers. In Culicover’s examples,

sentences involving negative inversion do not have this restriction of the deﬁniteness effect.

(86)
a. These are the books can’t nobody take home.
b. Before they invented them kicks, couldn’t nobody do that

c. Until Wu-tang came on the scene, didn’t nobody know ‘bout Inspector Dec.

(87)
d. * These are the book can’t Stacey take home.
e. * Before they invented them kicks, couldn’t John do that.

f. * Until Wu-tang came on the scene, didn’t the Source know ‘bout Inspector Dec.

Weldon argues that (88a) can either be an inversion construction or an existential

construction. Under the inversion analysis Weldon argues that (88a) means (88b); and as

an existential analysis, (88a) means (88c).

49

(88)

a. Ain’t nobody that’s happy rich.

b. Nobody that ‘s happy ain’t rich. (meaning: Nobody that’s happy is rich)

0. It/There ain’t nobody that’s happy rich. (meaning: There isn’t anybody who is happy
being rich.)

However, as far as I can tell, (88b) is not a possible interpretation for (88a). This
was not only my judgment, but the judgment of several other speakers of AAE as well. As
a sentence, (88b) is ungrammatical (shown below as 89b). The ambiguity that Weldon
states is predicted to exist under Weldon’s structural analysis because the sentence could be
existential (as in 89c or d) or an inversion construction (89b). There is no ambiguity in its
meaning. However, since these examples are not ambiguous, it is unlikely that (89a) can

ever have an inversion analysis as Weldon predicts.

(89)

a. Ain’t nobody that’s happy rich

b. * Nobody that’s happy ain’t rich.

c. It ain’t nobody that’s happy rich

d. There ain’t nobody that’s happy rich.

2.5. Summary

We have discussed the ﬂaws in the analysis of the NI constructions in AAE
proposed so far. What we have discovered from these studies is that when two analyses
for NI constructions are proposed, there is no account for why the deﬁniteness effect
inﬂuences the grammaticality of the sentence. When an inversion analysis is proposed, the

definiteness effect is generally ignored. As was illustrated above, if NI sentences pattern

50

like inversion sentences, then the deﬁniteness effect should not be a factor in determining
sentence grammaticality.

On the other hand, in claiming the NI sentence incorporates just an existential
structure, as Martin does, the problem that presents itself is how to account for why the
expletive does not show up in particular NI sentence constructions, especially those that do

not involve the copula be.

51

CHAPTER 3

With the exception of Martin, all the authors discussed in the previous chapters
conclude that NI sentences involve two distinct constructions, an inversion analysis and an
existential analysis. There are several reasons why these authors hold to this analysis.
First, not all NI sentences involve the use of existential be. Second, in NI sentences that
do not contain existential be , an overt expletive there cannot appear while in sentences
containing be an overt expletive can appear. Third, in existential sentences individual
level predicates cannot appear, but they do appear in NI constructions. Nonetheless, we
have seen that in all cases of NI, whether they involve existential be or not, the
deﬁniteness restriction is placed on their subjects. The definiteness restriction is not
predicted under an inversion analysis, since negative fronting in SE does not obey this
restriction.

In this chapter, I will argue, because of the definiteness effect, that all NI inversion
structures are like existential constructions, even in cases that do not involve existential be.
The main problem to be solved is the pattern of data in (90). Speciﬁc attention will be

given to sentences patterning like (90c), where the expletive cannot appear.

52

(90)

a. There ain’t nobody in the room.
b. Ain’t nobody in the room.

c. * There can’t nobody tag you.

d. Can’t nobody tag you.

I will argue that (90c) is ungrammatical because the post-auxiliary subject nobody
cannot receive Case, while in (90a) it can. Because of this, the pre-auxiliary subject

position can be ﬁlled in (90a) but not in (90c).

3.1 . Case Theory

Before explaining why the expletive there cannot be overt in certain NI sentences,
I will ﬁrst discuss why it is in general that Noun Phrases (NP)s need to be Case assigned.
Like all NPs we expect the expletive there to meet the same Case requirements as do other
N Ps. Descriptively, Case is assigned to an NP in a structural position that relates to a
Grammatical Function (GF) in a sentence (Chomsky 1981). Grammatical Functions are
Subjects, Direct Objects (DO), Indirect Objects (IO), and the object of the preposition (OP).
There are two different types of Case in English: Nominative Case and Accusative.

Case can be assigned to the NP by either a Verb (V) or a preposition (P),

however, N(ouns) and A(djectives) can not assign Case to an NP as shown in (91).

(91)

a. NP: * the destruction the city

b. VP: destroy the City

c. AP: * destructive her self-esteem

(1. PP: with the boy

53

In (91a) , the N destruction does not assign Case to the NP the city, which
violates the Case Filter. The Case Filter speciﬁes that every NP with a phonetic matrix
must receive Case. With the Case Filter, we can account for why (91a) is ungrammatical.
The same goes for AP; when an A takes as its argument an NP the result is ungrammatical,
as shown in (91c). However, when either V or P take on an NP argument the result is
grammatical. We can account for the difference between (91a and c) and (91b and d) with
Case assignment.

In cases where the NP argument is ungrammatical, we notice that neither a verb nor
a preposition appears. In the same NP and AP construction found in (91a and c), the NP

argument can appear when the preposition of is added:

(92)
a. NP: the destruction [of] the city

b. AP: destructive [of] her self-esteem

From this we can conclude that the difference between NPs and APs and VPs and PPs is
that Vs and Ps can take on NP arguments and Ns and As cannot. The difference that
allows the NP argument in VP and PP is Case.

N ow that we know that all NPs occuning in a sentence need to be assigned Case,
we can assume that the NP subject of a clause needs Case as well. Under the Minimalist
theory (Chomsky 1993), Case assignment is accomplished through raising. It is presumed
that the subject and the object start in VP where the theta- roles are assigned. Both then
raise to their respective AGR phrases. This movement allows all structural cases to be
assigned under Spec head agreement.

Consider the sentence (93):

(93) John met Bill.

54

The basic clausal structure is assumed to be (95). N ominative Case is assigned in Spec
AGRSP by T + AGR and Accusative Case is assigned in Spec AGROP by V + AGR.

(94)

In order to derive the correct word order from (94), we must assume that
movement of the subject is overt. The subject raises from Spec VP to the Spec of AGRS. T
raises from the head of TP to the head of AGRs where it forms a Spec head relation with
Spec AGRS.5 Although the Object must raise from the complement position in VP to the
Spec of AGRO, here movement is covert. The verb also moves covertly to the head of
AGRO forming a Spec-head relation with the object

The fact that subject movement is overt and object movement is covert can be

shown by the examples in (95), from Lasnik (1993).

 

5 Note that Martin, Sells et al., as well as Weldon utilize IP. Instead of IP I utilize AGRSP and TP

which are both the equivalent 0f IP. IP is the conﬂation of two separate heads AGRSP and TP. (See Pollock (1989) and
Chomsky (I993).

55

(95)

a. * has been arrested John.

b. John has been arrested.

c. The police have arrested John
(1. * The police have John arrested.

In (95a), the NP John stays internal to VP whose verb is transitive and the
sentence is ungrammatical. In (95b), assuming that John started in VP (the position that
John is in (95a)), subject raises out VP to Spec AGRS. In (95c) when movement of the
object is covert, the sentence is grammatical. This contrasts with (95d) which is
ungrammatical because the object has undergone overt movement to Spec AGRO. The
reason that subject N Ps must undergo overt movement is that English has a Strong NP
feature which requires the subject of a clause to be filled. This principle is called the
extended Projection Principle (EPP).

If not for the EPP then (95a) should be grammatical; the subject, like the object
could remain in VP and raise at LP to Spec AGRS. However, though Case could be
assigned by raising the subject at LP, the EPP requires the subject to raise at PF. This fact
is obvious in that (95b) is grammatical when overt movement of the subject takes place.

Furthermore, Lasnik shows that the subject does not move to AGRS to get Case. In
sentences involving the verbs like believe and consider which can take an infinitive
clause as their argument, Lasnik shows that the subjects of these clauses move to Spec
AGRS even though this position is without Case. This demonstrates that Subject moves to

AGRS to satisfy the EPP and not Case.

56

(96)
a. * I hope John to be out of trouble.
b. I consider John to be out of trouble.

c. I believe John to be a coward out of trouble.

In (96a), the sentence is ungrammatical because its embedded subject is not
assigned Case by the verb hope. On the other hand, the verbs consider and believe can
exceptionally assign accusative Case to the subject of their clausal arguments. Like
objects, the subjects of inﬁnitive clauses are assigned Case by raising at LP to AGROP
where a Spec-head relation is established with the raising of the subject to Spec AGRO and
the verb to the head of AGRO.

This assumes that John, the subject of the inﬁnitive clause starts out inside the
predicate before overt movement to AGRSP occurs as shown in (97a). At PF, John
satisﬁes the EPP by raising to the subject position Spec AGRS as shown in (97b). Finally,
at LP, in (970) John raises to AGRO where it receives accusative Case from the verb

consider which also raises to the head of AGROP from VP.

(97)
a. ...VP b. AGRs
/\ /\
be V PP Johni TP
/\ /\
John DP P to T AGROP
/\ /\
P PP VP
out of trouble /\
be V V’
/\
t,- PP

out of trouble

57

/\
t2 AGROP
/\
J0hl'l3 AGRo’

consider4 VP

/\
t4 AGRSP

These kinds of structures (96b and c) are known as Exceptional Case Marking
(ECM) structures. They are called ECM because accusative Case is being assigned to a

subject and not an object
(98)

a. I believe John

b. I believe John to be a fool.

58

In (98), the arguments of the verb believe are quite different. In (98a) John is the

complement of V at PF, while John in (98b) is not. It is in the Spec Position of AGRS in

the inﬁnitive Clause at PF. In (98b), the entire inﬁnitive clause is the complement of V and

not just John. However, at LF John in both (98a and b) must raise to the Spec of AGRO

in order to be assigned Case as shown in (99). (Only relevant parts of the structure are

shown.)
(99)
a. Spellout LF
...VP AGROP
/\ /\
I; V, John; AGRo’
believe John believel VP
/\
V’
/\
t1 12
b. Spellout LF
...VP AGROP
/\ /\
V’ Johnz AGRO’
./\ . /\
beheve AGRgP believe] VP
/\
John to e a fool V’
/\
t1 AGRSP

t; to be a fool

3.2. Case and Expletives
Now given the fact that all NPs need Case and that existential subject are also NPs, it '

follows that they also need Case. Lasnik (1995) along with Chomsky (1993) shows that

59

expletives in existential sentences occupy the position of the subject. The expletive, though
it is not assigned a theta-role, meets all the morphological requirements of the Subject
ﬁlling the requirements of tense and agreement. The associate (the post-auxiliary subject)

is assigned partitive Case and its theta-role is assigned by the predicate.

(100)
a. A strange man is t in the garden.
b. e is a strange man in the garden.

c. There is a strange man in the garden.

In (204a) Lasnik (1995) argues (revising Chomsky’s (1993) assumption) that overt
movement of the subject does not takes-place in order for the NP subject a strange man
to receive Case. Under Lasnik’s theory, the subject originally receives Case (partitive
Case) where it is positioned in (100b). Lasnik argues that it would not have to move until
LF and that the overt movement in (100a) occurs only because in (100b) the empty subject
position violates the EPP. In other words, overt movement occurs in order to meet the
strong NP feature of English. Furthermore, in (100C), Lasnik argues that when an
expletive is added, the post-auxiliary subject (the associate) can remain in the post-auxiliary
position provided the expletive checks the Strong NP features and that its associate receives
partitive Case. Lasnik concludes that both the expletive and its associate must be assigned
Case independently. When either one of the two is not assigned Case the result is

ungrammatical as shown in (101).

60

(101)
a. I want there to be someone here at 6:00.
b. * I want there someone here at 6:00.

c. * I hope there to be someone here at 6:00.

The contrast between (101a) and (101b) shows that be is required to assign Case to
the associate. The contrast between (101a) and (101c) shows that the expletive also needs
Case: want is able to assign accusative Case to the expletive, but hope is not, just as in
(96) above.

Not all verbs can assign partitive Case to the associate however, as the data in (102)

shows:

(102)

a. Someone laughed

b. * There someone laughed.
c. Someone arrived.

d. There arrived someone.

Intransitive verbs like laugh in (102a and b) where the subject is generated in Spec VP
cannot assign partitive Case. This is why the expletive cannot appear in (102d).
Unaccusative verbs like arrive whose argument is internal to VP as in (102c and d)
can assign partitive Case. Lasnik argues that the associate in (102d) is assigned partitive
Case while the expletive is assigned nominative Case. This establishes the fact that the

expletive and its associate are assigned Case independently.

3.3. Case and N I constructions
Now that we have established that the expletive and its associate must meet

requirements of Case independently, we can now argue that the expletive cannot appear in

61

 

certain NI sentences since its appearance blocks the associate from raising at LP to get
Case.

The general structure of the NI construction is shown in (103).

(103)

In (103), the negative auxiliary is in the head of TP. XP can be either NP, PP, AP or VP.
For NI sentences given either an “inversion” analysis or an “existential” analysis, their

structures are almost identical and can be both represented by the trees in (103) as (104)

illustrates.
( 104)
a. TP
/\
T VP
/\
V’
/\
V PP
ain’t
nobody P’
P NP
in
the room

62

b . AGRSP

AGRS’
TP
/\
T VP
can’t
nobody V’
V NP
tag you

Returning to the earlier argument, we learned that an inversion analysis would assume that
the auxiliary raises above the subject and assigns Case through its trace to the subject.
However, the inversion analysis ignores the fact that NI sentences demand their subjects to
be indeﬁnite. On the other hand, under an existential analysis all subjects which are

internal to VP must be indeﬁnite. This incorporates the deﬁniteness effect.

(105)
a. There is a man in the room.
b. * There is the man in the room.

c. A/The man is in the room.

(106)
a. There arrived a man.
b. * There arrived the man.

c. We man arrived.

Belletti (1988) shows that when the subject stays in VP that it must be indeﬁnite as

shown in (105a and 106a). When in Spec AGRS the indeﬁniteness restriction does not

63

affect the subject, which can be either indefinite or deﬁnite, as shown in (105c and 106c).
This generalization holds true whether the verb is existential be or not as (106) shows.
As mentioned earlier in Chapter 2, the inversion analysis is based solely on the fact

that the expletive cannot occur in sentences where the existential he does not occur.

(107)

a. * There can’t nobody tag you.

b. * There don’t nobody play with me.
c. Can’t nobody tag you.

C]. Don’t nobody play with me.

e. Ain’t nobody home yet

f. There Ain’t nobody home yet.

It was earlier determined that since in SE N I sentences are not restricted to
indeﬁnite subjects, an inversion analysis for N I in AAE was ruled out. First in view of
Belletti’s data, the structures in (104) predict that all subjects in NT sentences will be

indeﬁnite, a shown in (108).

(108)

a. Ain’t nothing going on.

b. * Aint’ the party going on.
c. Won’t nothing stop us now.

d. * Won’t John stop us now.
Second, taking into account Lasnik’s argument that the expletive and its associate

must be assigned Case independently, we know that if the expletive is present, there must

be a way of assigning partitive Case to the associate. We have already seen that be is able

64

 

to assign partitive Case. Modal auxiliaries, on the other hand, can only take VP arguments,

as shown in (109).

(109)

a. John can’t see. VP

b. * John can’t to see. [P
c. * John can’t a man NP

(1. * John can’t in the room. PP

It follows from this that modal auxiliaries do not assign partitive Case. We now
have an explanation for the data in (107 ). Further evidence that supports this analysis
comes from the following data. There are cases where a NT sentence involves be and the
expletive is still not allowed to appear. That is even when be appears, sometimes its

presence is not enough to allow the expletive to appear.

(110)

a. Can’t nobody be in the room.

b. Can’t be nobody in the room.

c. * There can’t nobody be in the room.

d. There can’t be nobody in the room.

In (1100 and (I), there is a slight difference which determines whether the expletive
can appear or not In (110C), when be is to the right of the post-auxiliary subject, the
sentence is ungrammatical, while on the other hand, (110d) is grammatical when be is to
the left of the subject This seems to suggest that what allows the expletive to appear

depends on whether be is positioned to the left or to the right of subject

65

Taking into account Case assignment, in (110d), we can assume that can’t is the

head of AGRS and with there in Spec AGRS it has a Spec-head relation with it Also, the
fact that nobody is internal says that the associate is assigned partitive Case. The same
should hold true for (110c). Can’t being the negative auxiliary is the head of AGRSP
which is in Spec-head agreement with the expletive there in Spec AGRS where Case is
assigned. However, for (l 10c), nobody is to the left of be. Though the subject is still
internal to VP, it appears that can’t is not a Case assigning verb. On the other hand, if be
were assigning Case to nobody, (110c) should be grammatical given that both the
expletive NP and the NP nobody would be assigned Case independently. Since the be in
(110c) does not assign Case to nobody we can assume that nobody like the subject of
inﬁnitive clauses must raise at LF to be assigned Case.

However, unlike subjects of infinitive clauses, the associate does not raise to
AGRO, nor it is in the Spec AGRS of an embedded clause. In existential sentences, the
associate is in complement position of VP. Unlike the subjects of inﬁnitive clauses, the
associate, because it is not in a nominative position, cannot raise to AGRO. If the associate
were to raise to AGRO there would be no way for an object to get Case. What distinguishes
the associate is the fact that it is assigned partitive Case. Partitive Case is a lexical property
which is only assigned by certain verbs and under govemment by the Case assigning head.

This explains the contrast between (110c and d) which appear here as (111).

(111)
a. * There can’t nobody be in the room.

b. There can’t be nobody in the room.

Munn (1996) has argued, that head government is still needed in minimalism.
Therefore, the simplest way to account for partitive is by government Though

minimalism has eliminated government, I will assume that something like government is

66

needed to assign partitive Case. Alternatively, I assume that partitive Case is assigned by

LF incorporation of the head of DP into the Case assigning verb as illustrated in (112).

(112) a. SpellOut LF
VP VP
arrive V DP V DP
/\ /\
D NP D V t, NP
a man a anive man
b. Spellout LF
VP VP
/\ /\
be V PP V PP
nobody DP P’ NoD Vno DP P’
P DP t,- NP P DP
in the room in the room

In (112a and b), at PF the verb takes DP as an argument At LP the incorporation is
established by the raising of the determiner head of DP to the Case assigning verb which is
the head of VP. This incorporation allows partitive Case to be assigned to the associate.

Given the fact that modals can’t assign partitive Case, Case cannot be assigned to
nobody in (110c). However, the fact that nobody is internal to VP but is not governed
by be indicates that it is not assigned partitive Case.

Based on Lasnik’s argument that the expletive and its associate need Case, I
conclude that (110c) is ungrammatical because the associate cannot receive partitive Case
from the verb be . In sentences where the associate is preceded by the negative modal,
partitive Case cannot be assigned from the modal verb nor from the verb be which follows

it as shown in (110c). In order to receive Case, the associate must move to the Spec of

AGIQ at LF. However, when the expletive appears, as in (1 10c), AGRs is ﬁlled and Case

67

has already been assigned to it. Since there already occupies Spec AGRs, the associate
cannot move at LF.

In summary, the reason why expletives cannot appear in NI sentences involving
modals stands on the fact that modals do not assign partitive Case. Like subjects of an
infinitive clause, these associates must raise at LF in order to be assigned Case. However,
when the expletive appears in the Spec of AGRS, it ﬁlls the only available Case position

that the associate can raise to.

3.4. Individual Level predicates in NI Sentences

Above, I have argued the expletive cannot appear in certain NI sentences because it
occupies the only position where its associate must raise in order to get Case at LF. In NI
sentences involving Individual Level predicates, I will argue that the same condition
applies.

Weldon argues for inversion in instances where the post auxiliary subject is an
individual level predicate. She bases her argument on the fact that the expletive cannot

appear in NI sentences containing individual level predicates.

(113)
a. Ain’t nobody a man

b. * There ain’t nobody a man.

However, there is evidence that from Spanish and Brazilian Portuguese which
suggests that individual level predicates need Case. Schmitt (1996) shows that ser is a
copula verb which is equivalent to the English copula be that takes on individual level
predicates, while on the other hand the copula estar chooses a stage level predicate.
Furthermore, Schmitt shows that individual level predicates are [+N] small clauses which

differ from stage level predicates which are [+V] small clauses. Schmitt argues that [+V]

68

clauses and [+N] clauses assign Case to their arguments in different ways. Schmitt argues
that since estar, unlike ser, only subcategorizes for AP and PP but not NP, that

individual level predicates need to be assigned Case as (114) shows.

(114)

a. Pedro é cozinheiro ser
Pedro is (a) cook. (That is his profession)

b. * Pedro esta cozinheiro. estar
Pedro is cook.
c. Pedro esta de cozinheiro. estar

Pedro is of cook.
(Pedro is playing the role of a cook today because the real one didn’t come)

(Schmitt 1996 p. 412 ex. 72)

Following Schmitt’s argument, in N I constructions involving individual level
predicates, Case must be assigned to both the predicate and the subject NP. Therefore, as
in the above cases, when the expletive appears in NI constructions, it occupies the only
Case position that the individual level predicate can raise to. Thus, in NI sentences
containing individual level predicates, the expletive cannot appear because either the

associate or the individual level predicate, must raise at LP to be assigned Case.

3.5 Summary

In this chapter, I have shown that expletives in AAE NI sentences containing
modals cannot appear due to Case. In NI sentences containing modal auxiliaries and
Individual level predicates, I argue that when the expletive appears it blocks the associate
and the individual level predicate from raising at LP to get Case. I ﬁrst showed that NPs in
general need Case and that the expletive and its associate are assigned Case. by independent
Case assigners. Second, I showed that modal auxiliaries cannot take on NP arguments. In

addition, the fact that NI sentences must obey the deﬁniteness restriction provides further

69

evidence that the negative auxiliary does not undergo inversion. Third, I conclude that the
associate that is preceded by a modal auxiliary is not assigned Case since the modal cannot
take on NP arguments. Last, I argue that the associate must raise at LP in order to be
assigned Case, and when the expletive appears movement is blocked.

In NI sentences involving Individual Level predicates, I show following Schmitt
(1996) that it also lacks Case. Like the associate that is preceded by a modal verb, the
Individual Level Predicate must also raise at LP to be assigned Case. When the expletive

appears in these cases, it blocks movement at LF.

70

 

CONCLUSION

Looking at the work done previously on AAE NI sentences, the debate has gone
back and forth in determining if two analyses are needed . The two competing theories
base their arguments on whether an overt expletive can appear or not, and on the fact that
all N I sentences adhere to deﬁniteness restriction. Those supporting that two analyses are
needed argue that inversion occur in instances where the negative auxiliary is a modal and
the expletive cannot appear. Supporters of a single analysis hold on to the fact that the
deﬁniteness restriction rule requires its subject to be indeﬁnite.

When considering the two approaches, it has become clear that one would have to
ignore either the deﬁniteness restriction or that fact that expletives cannot appear in
sentences with modal auxiliaries. As a result, Labov, Sells et al., and Weldon, argue that
N I sentences require two analyses, but ignore the effects of the deﬁniteness restriction.
Martin argues that NI sentences are existential due to deﬁniteness restriction, and ignores
the fact that the expletive does not appear in NI sentences containing modals.

I have proposed an analysis that encompasses both the deﬁniteness restriction and
non-expletive sentences. NI sentence are not inversion constructions because NI sentences
of SE do not follow the deﬁniteness effect I also showed that if inversion truly occurs in
NI sentences, it is ungrammatical when embedded is wh-clauses. This analysis argues that

what enables the expletive to appear is determined by be’s ability to assign Case to the
post-auxiliary subject The examples demonstrate that existential be is generated in a
position where it is either preceded by the initial auxiliary or by the post-auxiliary subject.
Furthermore it was determined that when be governed the post-auxiliary subject the
expletive could appear. This discovery then lead to the fact that when he follows the post-

auxiliary subject, it is not enough to allow the expletive to appear.

71

Lasnik’s study on existential sentences, illustrates that the expletive subject and its
associate are assigned Case independently. The auxiliary assigns nominative Case to the
expletive subject and be or an unaccusative verb assigns partitive Case to its associate. As
a result, at PF the expletive and its associate have Case assignments that are independent of
each other. Therefore in NI constructions, I conclude that the expletive cannot appear in
cases involving a modal since modals cannot take on NPs and they cannot assign partitive
Case. Therefore, the associate cannot raises at LP to be assigned Case and when the E-

expletive appears. This allows all NI constructions to be given a single analysis.

 

72

REFERENCES

73

REFERENCES

Belletti, Adriana. (1988) The Case of Unaccusatives, Linguistic Inquiry, 19: 1-34.
Chomsky, Noam. (1981) Lectures on Government and Binding, Foris, Dordrecht

Chomsky, Noam. (1993) A Minimalist Program for Linguistic Theory, in K. Hale and
SJ. Keyser eds. The view from building 20, 1-52, MIT Press, Cambridge, Mass.

Culicover, Peter W. (1991) Topicalization, in version, and complementizers in
English, in D. Delﬁtto, M. Rveraert, and f. Stuurman. eds., Going Romance and
Beyond 1-43, OTS Working Papers, University of Utrecht, Utrecht.

Klima, E. (1966) Negation in English, in J. Fodor and J. Katz, eds., The structure of
Language (1967), Prentice-Hall, Englewood Cliffs, New Jersey.

Labov, William (1972) Negative Attraction an Negative Concord in English Grammar,
Language , 48: 773-818.

Lasnik, Howard. (1993) “Lectures on Minimalist Syntax”, MIT Occasional Papers in
Linguistics #1, Cambridge, Massachusetts.

Lasnik, Howard. (1995) Case and Expletives Revisited: On Greed and Other Human
Failings, Linguistic Inquiry , 4: 615-633.

Lasnik, Howard. and Saito, Mamoru. (1991) On the Subject of Inﬁnitives, in Papers
from the Regional Meeting, Chicago Linguistic Society, University of Chicago,
Chicago.

Martin, Stefan. (1992) Topics in the Syntax of Nonstandard English, Doctoral
dissertation, University of Maryland, Collage Park.

Martin, Stefan. (1993) ‘Negative Inversion’ Sentences in Southern White English
Vernacular and Black English Vernacular, University of Maryland Working
Papers :49-56.

Milsark, G. L. (1977) Toward an Explanation of Certain Peculiarities of the Existential
Construction in English, Linguistic Analysis, 3: 1-29.

Munn, Alan. (1996) First conjunct agreement without government, In A. Alexiadou,
N. Fuhrhop, P. Law and S. Loehken (eds) ZAS Papers in Linguistics, 5, 85-104.

Pollock, Jean-Yves. (1989) Verb Movement, Universal Grammar, and the Structure of
IP, Linguistic Inquiry 3: 356-424

Schmitt, Cristina. (1996) Aspect and the Syntax of Noun Phrases, Doctoral dissertation,
University of Maryland, College Park.

Sells, Peter, John Rickford and Thomas Wasow (1996) An Optimality Theoretic
Approach to Variation in Negative Inversion in AAVE, Natural Language and
Linguistic Theory 14, 591-627.

74

Weldon, Tracy. (1997) Another Look at Negative Inversion in African-American
Vernacular English: An HPSG Account, MS., Ohio State University,
Columbus, Ohio.

75

 

"1111111111111155