Jespersen’s Cycle and the scope of negation in American Sign Language
Abstract. The goal of the present study is to contribute novel evidence in support of the idea that
some languages have a lexicalized distinction between external and internal negation. We
present data from American Sign Language (ASL) and a number of semantic scope diagnostic
tests that show that ASL has clearly split negation between a negative headshake and a manual
negation marker as instances of external and internal negation respectively. This analysis is
significant because recent research (Pfau 2015) has suggested that ASL, among other sign
languages, underwent a process akin to Jespersen’s Cycle, resulting in the predicted eventual
disappearance of manual negation as it was replaced as the head of NegP by the negative
headshake. Instead, our results show that this is not the case; rather than disappearing from the
language altogether, manual negation has undergone a semantic reanalysis and semantic
restructuring to become a marker of internal negation.
Keywords. American Sign Language, Jespersen’s Cycle, semantic restructuring, internal vs.
external negation, negative headshake, manual negation
1. Introduction
Negation is one of the more widely studied areas in the field of semantics. Patterns of negation
vary widely from language to language, leading authors to attempt to come up with larger
generalizations on the behavior of negation that occasionally overlook information from other
languages. Throughout the semantic literature, there has been a longstanding debate over the
distinction between external (sentential) and internal (constituent) negation, with a number of
contemporary logicians making the claim that there are no known languages that lexicalize the
distinction between the two (Gazdar, 1979; Levinson, 1983; Horn, 1989, a.o.). The difference
between external and internal negation for many logicians lies in syntactic scope rather than in
lexical distinction. External negation has been claimed to take scope in a functional projection
located higher in the syntax than Tense Phrase (TP) (Laka, 1990/1994; Cormack & Smith, 2002;
1
Poletto & Zanuttini, 2013). Internal negation, on the other hand, has been argued to take scope in
a low Focus Phrase (FP) (Belletti, 2004; Jayaseelan, 2001; Butler, 2003; Jayaseelan, 2008;
Kandybowicz, 2013). External negation is illustrated in English in (1a), where the overall
polarity of the proposition is negative and the predicate is denied. Example (1b), on the other
hand, shows that would be considered internal negation. In this proposition, the overall polarity
remains positive, with a negative predicate being affirmed by the subject.
(1) a. John does not regret having read the book.
b. John regrets not having read the book.
For languages like English, this generalization that languages mark external and internal
negation using the same lexical item seems to hold; both external and internal negation are
usually marked through the lexical item not. English’s negative prefix comes as an exception to
this generalization, as it is often treated as internal negation (Horn, 1989). Even analyses that
acknowledge this lexical distinction in English, however, do not always acknowledge a
fundamental difference between external and internal negation, and, instead, tend to assume the
existence of a common system underlying both (De Clercq, 2013). The interpretation of the
scope of negation instead depends largely on other factors such as prosodic stress (2) and
syntactic movement of the negation marker. However, there is a growing body of literature that
indicates that languages like Korean (Yoon, 2008) and Malay (Kroeger, 2014) use a different
marker for each type of negation. The distinct scope of negation is explicitly marked through the
negative lexical item that a proposition uses.
(2) a. We could not go to school (because it was a snow day).
b. We could NOT go to school (but we could go to the movies instead).
2
In the present paper, we add to that list by contributing novel evidence, on the basis of a number
of semantic scope diagnostics, that American Sign Language (ASL) exhibits two distinct patterns
of negation. Further, we argue that these distinct patterns arise as a result of the language
following Jespersen’s Cycle in a unique way.
2. Negation in ASL
ASL and other sign languages, unlike spoken languages, make use of non-manual markers to
impart real linguistic information. For ASL, these markers can take the form of eyebrow
furrowing for wh-questions, eyebrow raising for polar questions and topic marking, and a
headshake to signal negation, to name a few. Past studies on negation (Veinberg & Wilbur, 1990;
Anderson & Reilly, 1997) have revealed that ASL has several different options available to it in
terms of external negation. Stokoe (1960) first noted that a side-to-side headshake can be used to
signal negation on a syntactic level, regardless of the presence or absence of the sign commonly
glossed as ‘NOT.’ This general pattern is shown in (3), below, where the scope and duration of
the negative headshake is marked by the line over the gloss. In this construction, the sign ‘NOT’
is optional, and its inclusion or exclusion does not alter or contribute to the interpretation of
negation in the proposition. The negative headshake alone is capable of negating the entire
proposition.
________________________neg.
(3) WOMAN (NOT) GIVE-ME RED BOOK
‘The woman didn’t give me the red book.’
External negation in ASL allows for what appears to be an optional negative concord (NC)
construction involving both a negative headshake and a manual negation marker. In
3
constructions like (3), where there appears to be two instances of negation in a single clause, the
polarity of the proposition remains negative. Two negatives, in this case, do not make a positive.
But, unlike true NC languages, it is not the case that both negative markers are capable of
conveying logical negation. The manual negation marker, analogous to n-words in spoken
languages, does not convey logical negation; the facts outlined in (4a) and (4b) illustrate that the
pattern of NC in ASL is more similar to the pattern of negative spread seen in French than it is to
true NC languages. In French, ne does not contribute to logical negation. Instead, pas is the
propositional operator that negates the proposition, and, as a result, the ne…pas construction is
interpreted as being only a single instance of negation (Giannakidou, 2005). In ASL, logical
negation is expressed through the negative headshake, which explains the language’s previously
discussed ability to drop the manual negation marker in instances of external negation without
affecting meaning.
(4) a. *MARY NOT SEE PAUL
Intended meaning: ‘Mary didn’t see Paul.’
______________neg.
b. MARY NOT SEE PAUL
‘Mary didn’t see Paul.’
_________neg.
c. MARY SEE PAUL
‘Mary didn’t see Paul.’
The scope and duration of the negative headshake can vary according to a few set patterns. When
the headshake co-occurs with the pre-verbal sign ‘NOT,’ the headshake either accompanies only
‘NOT’ or may optionally spread to cover the entire verb phrase (5a). This optionality of
spreading is illustrated by the brackets in the non-manual line of the gloss. In contrast, when
‘NOT’ is absent, the headshake must obligatorily spread, as seen in (5b-c) (Neidle et al., 2000).
The final pattern of non-manual negation that ASL exhibits involves the headshake beginning
4
with the first sign of the proposition and obligatorily spreading over the entire tense phrase (5d).
The optionality of the manual negation marker in this type of construction does not affect either
the spread of the headshake or the grammaticality of the proposition itself; crucially, if the
headshake extends over the entire TP, the negation is not interpreted as negating the subject
noun. Instead, it is still the case that the predicate is negated.
____ (____________neg.)
(5) a. JOHN NOT BUY HOUSE
‘John is not buying a house.’
____________neg.
b. JOHN BUY HOUSE
‘John is not buying a house.’
_____neg.
c. *JOHN BUY HOUSE
‘John is not buying a house.’
_______________________neg.
d. JOHN (NOT) BUY HOUSE
‘John is not buying a house.’
Headshakes are also frequently used in spoken languages, but they do not hold the same
grammatical functions as they do in sign languages. This is due in part to the idea that the
primary function of a headshake for speakers of spoken languages is to provide emphasis or a
non-verbal no. However, they cannot be used alone to produce negation the way they can in sign
languages, as shown in (5b) for ASL, and they lack the sharp onset and offset that headshakes in
sign languages have been identified as possessing (Veinberg & Wilbur, 1990). Language contact
between spoken and signed languages has been claimed to be responsible for the adoption of
negative headshakes into sign languages where they eventually grammaticalized and became a
lexical part of negation (Pfau, 2015).
Because sign languages have been historically understudied, there have only been a few
studies on ASL negation that have looked at the possibility of internal negation as being separate
5
from external. Even though it is the most-studied and longest-studied sign language to date,
much of the research on negation in ASL has focused on descriptive aspects of the language
rather than on examining the theoretical implications for the facts from the language. There are,
of course, exceptions to this generalization. In 2004, Zeshan conducted a cross-linguistic
typological study of negation in signed languages with the idea that a negative headshake could
be used to mark internal negation. If a negative headshake could cover only a single word in a
proposition, it would be possible to argue that the scope of negation was limited to a constituent
rather than an entire proposition, though the results from the study showed that this is not the
case; the scope of the negative headshake cannot be limited to express internal negation for any
of the languages included in the study. Crucially, her study included LSQ, which is in the same
language family as ASL, but did not include ASL itself.
Fischer (2006) analyzed negative incorporation, a type of negation limited to a small set
of verbs where negation is marked through an outward twisting motion incorporated into the sign
itself, as a type of internal negation. According to her analysis, the negative forms of stative
predicates like ‘GOOD,’ ‘WANT,’ ‘LIKE,’ and ‘HAVE’ are derived through compounding the
predicate with ‘DON’T’; because by definition the negation is attached to a specific predicate
form, she argues that it is not truly acting on the entire proposition. She also notes the possibility
that ASL may use the manual negation marker ‘NOT’ or the negative headshake for internal
negation, though she does not provide an argument either for or against such a claim.
3. Jespersen’s Cycle in ASL
The studies that have examined the theoretical function of manual negation in ASL indicate that
the apparent connection between negation in French and in ASL goes beyond the two languages’
6
use of negative spread. Recent research (Pfau 2015) claims that signed languages as a
typological class underwent a process akin to Jespersen’s Cycle, for which French is the classic
and best known example. In Stage 1 of Jespersen’s Cycle, the sole obligatory negation marker
occupies the syntactic location Spec, NegP. Over time, but still in Stage 1, a secondary form of
negation is adopted into the language to act to reinforce or emphasize negation. Eventually as the
language develops, negation is reanalyzed and the initial negation marker, ne in the case of
French, moves to occupy Neg0 instead of Spec, NegP. The reinforcer, pas for French, then
becomes the main marker of negation and renews Spec, NegP rather than being part of the VP.
In modern colloquial French, negation has developed to the point that ne is optional and is
frequently omitted without affecting meaning. It is expected to eventually disappear altogether.
As a result, ne can no longer be said to be the head of the NegP. Figure 1 below shows the
development of negation in French.
Figure 1: Jespersen’s Cycle in French (Pfau 2015)
7
The mechanism that Pfau describes for Jespersen’s Cycle in signed languages is similar; based
on the different ways that negation is expressed across sign languages as a larger typological
class, he argues that negation was originally carried by manual negation. Headshakes were
introduced into signed languages later due to influence from co-speech gestures from other
languages. As negative headshakes were picked up through language contact with spoken
language communities, they first exhibited behavior similar to the function of pas and signaled
emphatic negation. In this first step of Jespersen’s Cycle, the manual negation marker is analyzed
as a negative adverb that occupies Spec, NegP. The reinforcing headshake acts as a lexical part
of the negative adverb rather than occupying its own syntactic location. In Stage 2, the
headshake is reanalyzed as Neg0 while the manual negation marker remains in Spec, NegP. This
change differs from the pattern that we see in French because the lexical item that originally
carried logical negation never becomes the head of the NegP. Instead, it remains a specifier
where Pfau (2015) notes that it is already only optionally expressed and predicts that it will
eventually disappear.
Figure 2: Jespersen’s Cycle in signed languages (Pfau 2015)
8
We argue, however, that this is not the case in ASL. Instead, the manual negation marker has
undergone a semantic and syntactic restructuring and found a new job as a marker of internal
negation. As the marker of internal negation, we propose that the manual negation is no longer
part of the NegP. Instead, it has become a negative adverb that modifies a generic XP, as shown
in Figure 3.
Figure 3: Internal negation in ASL
4. Current Proposal
Our current proposal has two main points. First, our data confirms that the headshake has been
reanalyzed and grammaticalized as the first step of a semantic restructuring, as defined by
Eckardt (2006). As a result, it is only capable of marking external negation, and has acquired the
grammatical status of a propositional operator, i.e. the element “headshake” of syntactic category
NEG with the semantic type <t, t>. Within the Montague’s Intensional Logic (IL), we can
represent the semantics of headshake as follows: the translation of headshake into the IL gives
rise to the addition of the propositional negative operator ¬, which means that the interpretation
of NEG S will be ¬S’ by function-argument application (Partee 2004):
(6) a. [[headshake]]
b. TR(S)
=
=
λpt.¬p
TR(headshake)(TR(S))
9
=
λpt.¬p (TR(S))
=
¬S’
Second, we propose that the job of the manual negation marker, contrary to common assumption,
was not merely weakened or nullified in this process of grammaticalization, but has become the
marker of internal negation. It has undergone a change from propositional operator to predicate
modifier in a process of reanalysis in the sense of Hopper and Traugott (1993). It was reanalyzed
as a negative adverb with a concomitant shift in meaning from a propositional operator of type
<t,t>, to a predicate modifier of various types, including <<e,t>,<e,t>> for intransitive verbs,
<<e,<e,t>>,<e,<e,t>>> for transitive verbs, etc. In a sentence with an intransitive verb, for
instance, the semantics of the manual negation marker, negating VP, can be defined as (a), and
its interpretation with VP follows the translation rule given in (b) (à la Partee 2004):
(7) a.
b.
[[MNVPi]]
TR(VP2)
=
=
λP<e,t>. λx. ¬P (x)
TR(MNVPi)(TR(VP1))
= λP<e,t>[λx[¬P (x)]](TR(VP1))
= λx[¬VP1’ (x)]
4.1 Internal vs. External Negation
ASL appears to exhibit two distinct patterns of negation. External negation, as shown in (8a),
requires the presence of a non-manual marker in the form of headshake negation. Internal
negation, shown in (8b), only requires the presence of the manual negation marker ‘NOT.’
__________________________________neg.
(8) a. MOTHER FUTURE (NOT) SEE CHILD
‘Mother will not see a child.’
b. MOTHER FUTURE NOT SEE CHILD
* ‘Mother will not see a child.’
‘Mother will not-see a child.’
10
Then the truth conditional meanings of above sentences can be informally represented as the
following, reflecting the relative scope of the negation and the quantifier:
(9) a. ‘Mother will not see a child.’: NEG [a child x [Mother will see x]]
b. ‘Mother will not-see a child.’: [a child x NEG [Mother will see x]]
Though it is easy to confuse the precise scope of negation, the results of the following tests
collectively support our proposal that the negative headshake and manual negation marker
‘NOT’ have clearly divided the labor as external and internal negation respectively. Among the
properties used to distinguish external from internal negation were: tag questions, the again-test,
the deliberately-test, metalinguistic negation, expletive negation, and try-to-V constructions.
4.2 Tag Questions
Tag questions were originally proposed as a method of determining wide from narrow scope
negation by Klima (1964). Positive or negative tag questions must be opposite in polarity the
sentence they attach to; because external negation, which has wide scope, results in a negative
proposition, the tag questions that attach to it must be positive, as in (10a). Internal negation,
which has a narrow scope, does not change the overall polarity of a sentence, so the polarity
remains positive. As a result, the tag questions that attach to instances of internal negation must
be negative like in (10b).
(10) a. It’s not late, is it?
b. John regrets not having read the book, doesn’t he?
Tag constructions are not common in ASL. Sentence-final ‘doubling’ of a sign where the
doubled sign is separated from the rest of the phrase by a pause has been analyzed as a type of
11
tag (Wilbur, 1999), but signers do not make frequent use of tag questions. However, though they
viewed the constructions as somewhat marked across the board, judgments from consultants
show the same pattern of tag question use outlined by Klima in ASL. In (11a), where the
presence of the negative headshake creates external negation, the positive tag question ‘TRUE
BUSINESS,’ which is roughly equivalent to ‘is it,’ is acceptable. When negation is also extended
to the tag question, as in example (11b), the resulting sentence was judged ungrammatical. These
results follow Klima’s pattern of tag question use because the tag question in (11b) is not
opposite in polarity the sentence it attaches to.
__________neg.
(11) a. TIME LATE,
‘It’s not late, is it?’
__________neg.
b. *TIME LATE,
‘It’s not late, isn’t it?’
_______________y/n
TRUE BUSINESS
_neg _______________y/n
NOT TRUE BUSINESS
In (12a) we show the manual negation marker expressing internal negation and resulting in an
affirmative proposition. The tag question that attaches to the proposition must be negative in
order to be grammatical, which is illustrated by the ungrammaticality of (12b), where both the
question and its tag are positive.
_neg _______________y/n
(12) a. JOHN SORRY NOT READ BOOK,
NOT TRUE BUSINESS
‘John regrets not having read the book, doesn’t he?’
_______________y/n
b. *JOHN SORRY NOT READ BOOK,
TRUE BUSINESS
‘John regrets not having read the book, does he?’
4.3 Again-test
12
The again-test was originally proposed as a diagnostic for object position, but as per Beck and
Snyder (2001), it has also been suggested to indicate the syntactic domain of negation. When
again is added to an affirmative sentence, the resulting interpretation is repetitive. Internal
negation shows identical behavior to affirmative statements and only allows for a repetitive
interpretation. External negation allows for both a repetitive and restitutive interpretation. In
example (14), the presence of external negation combined with the sign glossed as ‘AGAIN’
results in a repetitive meaning. But according to consultants, it also results in the additional
restitutive meaning shown below. This contrasts with the manual negation marker in (15), where
only the repetitive meaning is available. These results suggest that (15) is an example of internal
negation because it follows the pattern for affirmative sentences and the expected pattern for
internal negation.
(13) SALLY PAINT DOOR BLUE AGAIN
‘Sally painted the door blue, and she had painted the door blue before.’
_______________________________neg.
(14) SALLY (NOT) PAINT DOOR BLUE AGAIN
‘Sally didn’t paint the door blue again, but she had painted it blue before.’
‘Sally didn’t paint the door blue again, but it had been blue before.’
(15) SALLY NOT PAINT DOOR BLUE AGAIN
‘Sally didn’t paint the door blue again, but she had painted it blue before.’
4.4 Deliberately-test
Several different authors have proposed the deliberately-test as a diagnostic for the scope of
negation (Klima, 1964; Payne, 1985; McCawley, 1988). The introduction of deliberately to
instances of external negation necessarily results in ambiguity. Under one interpretation of (16),
the deliberateness of the act is outside the scope of negation; John intentionally avoided Mary. In
another interpretation, the deliberateness falls under the scope of negation, meaning that John did
see Mary but that it was not on purpose. In (17), however, where internal negation is expressed
13
through manual negation, ‘ON-PURPOSE’ is not negated and only the interpretation that John
avoided Mary on purpose is allowed.
____________________neg.
(16) JOHN MEANING SEE MARY
‘John avoided (did not see) Mary on purpose.’
‘John saw Mary, but not on purpose.’
(17) JOHN MEANING NOT SEE MARY
‘John avoided (did not see) Mary on purpose.’
4.5 Metalinguistic Negation
The term metalinguistic negation was first used by Horn (1985) to describe instances of negation
where what is negated is not the truth of a proposition, but what he called its assertability.
Examples in English include things like (18a-c).
(18) a. I’m not cold, I’m freezing.
b. It’s not Mrs., it’s Dr.
c. Joe didn’t cause some of the problems, he caused all of them.
Metalinguistic negation has also been identified as requiring a syntactic wide scope (Horn 1985,
Yoon 2008) and is only available to external negation. Because internal negation acts on
predicates rather than propositions, attempting to express metalinguistic negation through a
marker of internal negation would result in a semantically anomalous interpretation where what
is intended to be negated is instead affirmed. Though the syntactic structure of metalinguistic
negation in ASL is quite different from the structure in English, it is still the case that only
external negation can be used to express it. In example (19), we see the original proposition set
up as a yes/no question where the signer asks if they are happy. This is followed by negation in
the form of a negative headshake, and, finally, a corrected statement asserting the degree of the
signer’s happiness. (20), on the other hand, illustrates that the same idea cannot grammatically be
expressed with manual negation.
14
___________y/n _____neg.
(19) IX:1P HAPPY (NOT)
IX:1P THRILLED
‘I’m not happy, I’m ecstatic!’
____________y/n
(20) a. #IX:1P HAPPY NOT IX:1P THRILLED
‘I’m unhappy, I’m ecstatic!’
_________________y/n
b. #IX:1P NOT HAPPY
IX:1P THRILLED
‘I’m unhappy, I’m ecstatic!’
4.6 Expletive Negation
Portner and Zanuttini (2003) documented (21) from Paduan dialect of Italian, where the
sentence-initial negative marker is not apparently realized in the meaning. They propose that
these types of constructions create a comparative, where the likelihood of a true proposition is
judged against its negative alternative. If the likelihood of the true proposition is rated as being
less likely, negation is licensed to trigger an emphatic effect in expressing how unlikely or
unexpected the true proposition was.
(21) No
ga-lo
neg. has-s.cl
‘He ate everything!’
magnà tuto!
eaten everything
[Paduan Italian]
In ASL, expletive negation has never specifically been studied or documented, though the use of
headshake negation has been said to express uncertainty (McClave, 2001). McClave (2001) also
analyzed the use of headshake negation in (22) as being the result of behaviors picked up from
spoken English. Under this analysis, the negative headshake is not serving a linguistic or
grammatical function, and is, instead, more along the lines of a co-speech gesture. Given Pfau’s
(2015) analysis that headshakes have grammaticalized into signed languages and the relative
infrequency of non-meaningful co-speech gestures used by native signers, however, we argue
that (22) could also be analyzed as a type of expletive negation rather than a lateral headshake.
15
_____________________neg.
(22) WOW SHOW-UP MANY
‘Wow! Many (non-handed signs) showed up.’
(McClave, 2001; 55)
Like metalinguistic negation, expletive negation is only available to external negation (Yoon
2011). McClave (2001) reports that (22), with a headshake that we analyze as external negation,
is grammatical with the stated reading. In this interpretation, the negative headshake does not
impart any negative meaning on the sentence. For (23a) and (23b), consultants reported that
regardless of where the manual negation marker is placed, they were unable to construct a
proposition with the same meaning as (22). Instead, the sentences in (23) are semantically
nonsensical because consultants felt strongly that the negation should be expressed somewhere in
the meaning.
(23) a. #WOW NOT SHOW-UP MANY
‘Wow! Many (non-handed signs) showed up.’
b. #WOW SHOW-UP NOT MANY
‘Wow! Many (non-handed signs) showed up.’
4.7 Try-to-V Constructions
The final diagnostic test for the scope of negation that we present for ASL is the try-to-V
construction. This diagnostic is arguable, as its validity depends on the assumed syntax of try-to
constructions. However, their use as a diagnostic for the scope of negation stems from the idea
that in these types of constructions, narrow scope is only available to internal negation. Because
external negation results in a reversal of polarity for an entire proposition, negation with a
narrow scope cannot be expressed through external negation. Internal negation, conversely,
cannot be used to express wide scope negation because it is not capable of acting on a sentence
to reverse its polarity. This same pattern is seen in ASL, where a negative headshake is required
16
to express wide scope negation. (24a) and (24b) together show that the wide scope interpretation
is obligatorily expressed through non-manual headshake negation. (25a) and (25b), on the other
hand, show the opposite pattern. The narrow scope meaning can only be expressed through the
use of a manual negation marker. In (25b), even when the negative headshake only scopes over
‘LAUGH,’ the resulting sentence is judged ungrammatical.
_________________neg.
(24) a. BILL (NOT) TRY LAUGH
‘Bill didn’t try to laugh.’
b. *BILL NOT TRY LAUGH
‘Bill didn’t try to laugh.’
(25) a. BILL TRY NOT LAUGH
‘Bill tried not to laugh.’
_____________neg.
b. *BILL TRY (NOT) LAUGH
‘Bill tried not to laugh.’
From the above scope diagnostics, it is clear that ASL systematically differentiates between
manual and non-manual negation. Additionally, ASL follows the expected patterns for internal
and external negation with respect to the manual and non-manual markers respectively. The
negative headshake behaves the way we would expect external negation to in only licensing
positive tag questions, allowing for both a repetitive and restitutive interpretation when
combined with ‘AGAIN,’ creating ambiguity when used with ‘MEANING,’ licensing expletive
and metalinguistic negation, and only licensing a wide scope negation. Manual negation,
contrary to Pfau’s (2015) prediction, behaves the way we would expect internal negation to.
Manual negation licenses the use of negative tag questions, only allows for a repetitive reading
when combined with ‘AGAIN,’ does not result in ambiguity when used with ‘MEANING,’ does
not license either metalinguistic or expletive negation, and allows for a narrow scope. It should
17
be noted, however, that not all of the previously identified scope diagnostics for negation are
available to ASL; adverbial jumping and coordination do not offer the clear distinction between
negative headshake and manual negation that the previously discussed diagnostics provided. We
discuss these diagnostics in the following two sections, though we believe that the break in
pattern can be attributed to the syntactic structure preferences of the language rather than to
issues of semantic scope.
4.8 Adverbial Jumping
Not all diagnostics are available to ASL. Adverbial jumping, which involves negation of an
adverb under a semantic transformation known as “jumping,” does not seem to be a strategy in
ASL due to the language’s preference for expressing adverbial propositions like (26a) and (26b)
through rhetorical questions. Rhetorical questions are a commonly used construction in ASL and
function to provide emphasis, introduce new information, or establish a causal relationship
between two events. Like wh-questions and yes/no questions, rhetorical questions are associated
with a specific type of non-manual marking which is indicated in the gloss by the line over the
wh-sign.
_________________________neg. _____rhq
(26) a. IX:3P (NOT) WEAR JACKET
WHY
BECAUSE WEATHER
‘He doesn’t wear a jacket because of the weather.’
________________________rhq _________________________neg.
b. IX:3P WEAR JACKET WHY
(NOT) BECAUSE WEATHER
‘He wears a jacket not because of the weather (but for some other reason).’
Negation in (26a) and (26b) is external and carried by a negative headshake. This fact is not
altogether surprising when the syntactic structure of rhetorical questions in ASL is considered.
Each part of the statement is set up as an independent clause and negation is moved to the clause
18
where it is relevant to convey the desired meaning. In this case, it is the entire clause that is being
negated, which explains the licensing of external negation rather than internal.
4.9 Coordination
Sentential negation in coordinate structures should demonstrate a scopal ambiguity. ASL does
not show such an ambiguity, and instead arranges external negation to cover either the entire
proposition or just the negated clause. Internal negation does not play a significant role in
coordinate structures.
_________________________neg.
(27) a. JOHN DRINK ALCOHOL BUT (NOT) SMOKE CIGARETTE
‘John drinks alcohol, but doesn’t smoke cigarettes.’
__________________________________________neg.
b. JOHN (NOT) DRINK ALCOHOL, SMOKE CIGARETTE
‘John doesn’t drink alcohol or smoke cigarettes.’
______________________neg.
c. JOHN (NOT) DRINK ALCOHOL
BUT SMOKE CIGARETTE
‘John doesn’t drink alcohol, but he smokes cigarettes.’
The defined onset and offset of the negative headshake in ASL makes disambiguating exactly
what is being negated in coordinate structures much easier than in spoken languages. When the
headshake ends, it is a clear signal of how far the scope of negation extends; if it extends to cover
one VP of the coordinate structure, only that VP is negated and there is no ambiguity. Likewise,
if the headshake covers both VPs, both are negated and again there is no ambiguity.
5. Syntactic Scope of Negation: Evidence from Negative Polarity Items
External and internal negation have different syntactic structures; while external negation
occupies the Neg0 position of the NegP, our proposal for the structure of internal negation in
19
ASL does not involve the existence of a NegP. Instead, internal negation acts as a negative
adverb that modifies a NP or VP. The difference between the two structures is highlighted in
Figure 4 and Figure 5 below (taken from example 8).
Figure 4: External negation in ASL
Figure 5: Internal negation in ASL
Negative Polarity Items (NPIs) can offer further insight into the syntactic scope and location of
negation because they require the existence of a NegP to be licensed in a proposition. The sign
glossed as ‘ANY’ in ASL is sensitive to downward-entailing environments in the phrase ‘ANY
HEART-SOFT’ and cannot be licensed in positive polarity contexts (Schlenker, 2016). Though
the behavior of ‘ANY’ with respect to the full range of (non)veridical contexts remains to be
examined, Schlenker’s findings show that ‘ANY’ in ASL does show NPI behavior to some
extent. Building on these facts, examples (28a) and (28b) show that the sign is indeed sensitive to
negation outside of the fixed phrase in Schlenker’s study; where external negation is expressed
and the polarity of the proposition is negative, ‘ANY’ can be grammatically licensed. A
proposition with no negation expressed anywhere, like example (28b), is positive in polarity and
does not permit the use of ‘ANY.’ The most important example from this data set is (28c), which
shows that internal negation is also not sufficient to license the use of an NPI.
20
___________________neg.
(28) a. JOHN (NOT) SEE ANY+ONE
‘John didn’t see anyone.’
b. *JOHN SEE ANY+ONE
‘John saw anyone.’
c. *JOHN NOT SEE ANY+ONE
‘John didn’t see anyone.’
This generalization provides further evidence that the manual negation marker has become a
predicate modifier. According to the Negative Criterion (Zanuttini, 1991; Haegeman, 1995) NPIs
must move to Spec, NegP in the Logical Form (LF). External negation is located in the Neg0
position of the NegP. Internal negation’s occupation of Spec, NegP as predicted by Pfau’s (2015)
model is potentially problematic for the licensing of NPIs in ASL, since two separate lexical
items cannot occupy the same syntactic location at the same time. Our analysis that ‘NOT’ is
actually a part of a generic XP solves this problem, however, as it frees up Spec, NegP for the
NPI to move into. Additionally, it explains the ungrammaticality of (28c) because there is no
NegP created for the NPI to move into. Like English, ASL does not allow for subject NPIs. In
moving into Spec, NegP, the NPI ‘ANY+ONE’ moves into an A' position. Moving into Spec, IP
to occupy a subject position would require a disallowed movement from an A' to an A position.
21
It is also worth noting that ‘ANY’ can topicalize outside of the surface structure scope of
negation (30b), but this can be attributed to pre-movement licensing.
__________________neg.
(29) *ANY+ONE (NOT) SEE JOHN
‘Anyone didn’t see John.’
_____________________neg.
(30) a. BILL (NOT) BUY ANY BOOK
‘Bill didn’t buy any books.’
__________top
_______________neg.
b. ANY BOOK,
BILL (NOT) BUY
‘As for books, Bill didn’t buy any.’
6. Semantic Restructuring of Manual Negation
Historical evidence from sign languages is rare because prior to the 1960s, they were not
considered true languages and thus little effort was made to document or preserve data. The
development of video technology made preservation of sign language data more possible, but it
was not until the publication of Stokoe’s (1960) research that steps were taken to document and
investigate ASL on linguistic terms. In response to this fact, Pfau (2015) proposes that sign
languages as a typological class represent different points along Jespersen’s Cycle. For all sign
languages that have had their systems of negation analyzed, there exists a combination of both
manual and non-manual signs. The degree to which they rely on one over the other in expressing
external negation varies from language to language; signed languages that mainly rely on manual
negation, like Italian Sign Language (Lingua Italiana dei Segni, LIS) or Hong Kong Sign
Language (HKSL), represent the early stages of the cycle.
____neg.
(31) PAOLO CONTRACT SIGN NON
‘Paolo didn’t sign the contract.’
22
[LIS]
(Pfau, 2015; 25)
Languages that are commonly negated through the use of a headshake alone, like Flemish Sign
Language (Vlaamse Gebarentaal, VGT) or Indopakistani Sign Language (IPSL) are said to
represent a point further along in Jespersen’s Cycle. For these languages, just like for ASL, the
use of a manual negation marker is optional in expressing external negation because, according
to Pfau’s analysis, they have reached a point in Jespersen’s Cycle where the negative headshake
has reanalyzed and become the main marker of negation.
__________________neg.
(32) NOW GAME START
‘The game doesn’t start now.’
[VGT]
(Pfau, 2015; 29)
If we assume this to be correct, we can show that both the negative headshake and the manual
negation marker have undergone semantic restructuring since their initial roles. Semantic
restructuring is marked by an increase in cohesion and a decrease in weight and variability, as
characterized by Lehmann (1982, 1995). We propose that there have been three stages of
semantic restructuring in the scope of negation in ASL: (i) an initial semantic generalization
(bleaching) of the manual negation marker arose to remove the external negation meaning; (ii)
the negative headshake lost emphatic pragmatic meaning and picked up semantic meaning; and
(iii) the manual negation marker picked up new semantic meaning and became a predicate
modifier.
At the third stage of the semantic restructuring, the new meaning of the manual negation
as internal negation can be driven from its original meaning as external negation by a simple
type-shifting rule, for instance, for modifying an intransitive verb:
(33) [[ MNstage3 ]]
= λP. λx. [[ MNstage1 ]] ( P(x) )
= λP. λx. [ λp. ¬p] ( P(x) )
= λP. λx. ¬ P(x)
23
In order for a morpheme to undergo such semantic reduction or bleaching, it first has to lose its
intention. Rather than describing a specific idea, or even a limited set of ideas, it evolves and
begins to describe a larger range of things, eventually losing its meaning altogether (Haiman,
1991). It could be argued that manual negation in ASL undergoes a type of semantic bleaching
where it changes from serving a very specific purpose of negating entire propositions as a
propositional operator. As shown in (24) and (25) above, propositional operators cannot operate
on the level of individual constituents. They can only function with a wide scope on the level of
a larger clause. Predicate modifiers serve a much more general purpose and have the ability to
modify any potential XP.
In order to say that the manual negation marker ‘NOT’ in ASL has undergone a true
semantic restructuring, it must demonstrate a reduced structural scope, either syntactically or
semantically. Syntactically, the manual negation marker no longer has the potential to act as the
head of NegP, as seen by its inability to license NPIs. However, this inability to license NPIs also
demonstrates a reduced semantic scope. Based on Giannakidou’s (1999; cf. Zwart 1995)
definition of NPIs as being licensed by (non)veridicality, the grammaticality judgments from
(28), repeated in (34) below, show that manual negation does not reverse the veridicality of a
proposition. As a result, it cannot grammatically license NPIs. We attribute this fact to manual
negation having undergone a semantic type shift from <t,t> to <<e,t>,<e,t>>, etc., as noted
above. Anything of the semantic type <t,t> would be able to license NPIs because it would take a
truth value as its argument and return a truth value, meaning that it could affect the veridicality
of a proposition. The semantic type <<e,t>,<e,t>>, on the other hand, can only take a predicate as
its argument and can only return something of the same type. Manual negation results in a
24
sentence that is positive in polarity but that has a negative predicate, which would be expected
from the described type shift.
(34) *JOHN NOT SEE ANY+THING
‘John didn’t see anything.’
(35) *MARK NOT GO-TO STORE
‘Mark didn’t go to the store.’
The manual negation marker must also demonstrate a reduced syntagmatic variability; rather
than being able to move about freely in terms of its syntactic location, it should occupy a fixed
location relative to its function within the proposition. In ASL it is true that manual negation
cannot freely move around syntactically without affecting either meaning or grammaticality.
Instead, it occupies a pre-verbal position in which it is only permitted to appear immediately
before the constituent that it negates. The following examples show that moving ‘NOT’ to
different syntactic locations changes the intended meaning of the proposition.
(36) a. JOHN SORRY READ NOT BOOK
*Intended meaning: ‘John regrets not having read the book.’
Attested meaning: ‘John regrets having read not a book.’
b. NOT JOHN SORRY READ BOOK
*Intended meaning: ‘John regrets not having read the book.’
Attested meaning: ‘It is not-John that regrets having read the book.’
c. JOHN NOT SORRY READ BOOK
*Intended meaning: ‘John does not regret having read the book.’
Attested meaning: ‘It is the case that John is not sorry he read the book.’
d. *JOHN SORRY READ BOOK NOT
‘John regrets not having read the book.’
In comparison, the negative headshake exhibits more freedom with respect to syntagmatic
variability. Its scope can cover either a VP or an entire TP with no difference in meaning or
grammaticality. Where the examples in (36a)-(36c) show a discrepancy between their intended
meanings and their attested meanings, the examples in (37) show no such distinction. When the
25
onset of the headshake coincides with the beginning of the subject NP it is interpreted as being
external negation just as it would be if the onset came with the VP instead.
_______________________neg.
(37) a. DOG (NOT) CHASE CAT
‘The dog didn’t chase the cat.’
_________________neg.
b. DOG (NOT) CHASE CAT
‘The dog didn’t chase the cat.’
7. Support from Flemish
The traditional pattern of Jespersen’s Cycle results in a change where the original preverbal
element of negation eventually disappears from the language altogether in Stage III after a period
of semantic reanalysis. ASL presents a unique divergence from this pattern, but it is not the only
language for which a semantic reanalysis of the original negation marker has been proposed. All
of the West Germanic languages have undergone Jespersen’s Cycle, and almost all of them have
reached Stage III (Breitbarth & Haegeman, 2010). For Standard Dutch, this transition was
completed by the 17th century (Burridge, 1993), but Flemish varieties of Dutch have retained
both the original preverbal marker of negation, en, as well as the newer postverbal marker nie.
These facts have traditionally been accounted for through the analysis that Flemish represents a
point between Stage II and Stage III of Jespersen’s Cyle (Haegeman & Zanuttini, 1996; Zeijlstra,
2004).
(38) Valère (en) klaapt nie.
Valère neg. talks neg.
‘Valère doesn’t talk.’
[West Flemish]
(Breitbarth & Haegeman, 2010; 2)
Breitbarth and Haegeman (2010) proposed an alternative explanation for the pattern shown in
(38). Rather than having failed to reach Stage III, they argue that en in Flemish has reanalyzed as
26
a marker of emphatic polarity. In (39), a negative statement is contradicted by an emphatic
affirmative using en; under this view, Stage II represents an asymmetrical division of labor for
the two markers rather than an instance of bipartite negation. The two elements, en and niet, are
associated with different functional heads in a mechanism of reanalysis similar to Roberts and
Roussou’s (2003) upward reanalysis, and en is generalized to non-veridical contexts beyond
negation (Breitbarth & Haegeman, 2010).
(39) A: Hij zal niet komen.
he will not come
‘He won’t come.’
B: Hij/’t en doet.
he
EN does
‘No, he WILL.’
(Breitbarth & Haegeman, 2010; 8b)
The pattern of reanalysis post-Jespersen’s Cycle that is seen in West Flemish is not exactly
duplicated in ASL, but it does lend credence to the idea that an original negation marker can
reanalyze to find a new role after it loses its function as external negation.
8. Conclusion
We have shown that ASL follows Jespersen’s Cycle in a unique way. Just like pas in Jespersen’s
Cycle in French, a non-manual marker in the form of a negative headshake has been reanalyzed
to become the main carrier of external negation (Pfau 2015), which was confirmed by our scope
diagnostics. More importantly, we propose that manual negation has also undergone a semantic
and syntactic change, making it distinct from ne in French. Rather than becoming optional and
simply disappearing as the traditional pattern for Jespersen’s Cycle predicts, the results of our
scope diagnostics show that systematic asymmetries currently exist between the two negative
markers in ASL; the manual negation marker has undergone semantic reanalysis (Hopper and
27
Traugott 1993) as internal negation. Further, negation in ASL patterns with other paths of
meaning change involving semantic restructuring that can provide insight into the relationship
between diachronic change and synchronic meaning.
This newfound pattern in ASL is extremely interesting considering other reported
Jespersenian developments of the familiar kind which, for instance, have resulted in the original
negator (NEG1) being reanalyzed as a polarity-element/high Pol-head such as ne in French
(Rooryck 2008) and en- in West Flemish (Breitbarth & Haegeman 2013, 2014). This
development would involve upwards grammaticalisation (cf. Roberts & Roussou 2003). What
we find in ASL instead, however, is that a weakened or bleached NEG1 has taken on a new
function in the non-clausal domain, starting a new Jespersenian Cycle in an internal domain and
taking over that domain completely. Furthermore, the development pattern in ASL is important
when viewed in a wider negation context such as West Germanic cases of internal cycles which
reveals the opposite pattern such that it is NEG2 which is extended to the internal domain, not
NEG1, and another difference there is that the internal domain is primarily non-verbal (DPs, PPs,
etc.) (Van der Auwera & Neuckermans 2004; van der Auwera et al. 2006; Biberauer 2009,
2012).1
Acknowledgement. To follow.
1
We thank an anonymous reviewer of GLOW 39 for bringing our attention to the wider context of historical
development of negation.
28
References
Anderson, Diane and Reilly, Judy S. 1997. The puzzle of negation: How children move from
communicative to grammatical negation in ASL. Journal of Child Language 18.4, 411
429.
Beck, Singrid and Snyder, W. 2001. The resultative parameter and restitutive again. In Caroline
Ferry & Wolfgang Sternefeld (Eds.) Audiatur Vox Sapientiae: A Festschrift for Arnim
von Stechow, 48-69. Berlin: Akademie Verlag.
Belletti, Adriana. 2004. Aspects of the low IP area. In Luigi Rizzi (Ed.) The Structure of IP and
CP: the cartography of syntactic structures, volume 2. Oxford: Oxford Univeristy Press.
16-51.
Biberauer, Theresa. 2012. Competing reinforcements: when languages opt out of Jespersen’s
Cycle. In A. Van Kemenade & N. de Haas (Eds.). Historical Linguistics 2009: Selected
papers from the 19th International Conference on Historical Linguistics. Amsterdam:
John Benjamins, 3-30.
Biberauer, Theresa. 2009. Jespersen off course? The case of contemporary Afrikaans negation.
In E. van Gelderen (Ed.) Cyclical Change. Amsterdam: John Benjamins, 91-130.
Burridge, Kate. 1993. Syntactic Change in Germanic: Aspects of Language Change in Germanic
with Particular Reference to Middle Dutch. Benjamins: Amsterdam.
Butler, Jonny. 2003. A minimalist treatment of modality. Lingua 113: 967-996.
Breitbarth, Anne and Liliane Haegeman. 2014. The distribution of preverbal en in (West)
Flemish: Syntactic and interpretive properties. Lingua 147, 69-86.
Breitbarth, Anne and Liliane Haegeman. 2010. Continuity is change: the long tail of Jespersen’s
cycle in Flemish. In Anne Breitbarth, Christopher Lucas, Sheila Watts, and David Willis
(Eds.) Continuity and Change in Grammar. John Benjamins: Amsterdam. 61-79.
Cormack, Annabel and Neil Smith. 2002. Modals and negation in English. In Sjef Barbiers,
Frits Beukema, and William van der Wurf (Eds.) Modality and its Interaction with the
verbal system. Amsterdam: John Benjamins. 133-163.
De Clercq, Karen. 2013. A Unified Syntax of Negation. Ph.D. thesis. Universiteit Gent.
Deo, Ashwini 2015. Diachronic semantics. Annual review of Linguistics 1, 179-197.
Eckardt, Regine 2006. Meaning Change in Grammaticalization. An enquiry into Semantic
reanalysis. Oxford: Oxford University Press.
Fischer, Susan D. 2006. Questions and negation in American Sign Language. In Zeshan Ulrike
(Ed.) Interrogatives and Negative Constructions in Sign Languages 165-197.
Amsterdam: Ishara Press.
Gazdar, Gerald 1979. Pragmatics: Implicature, presupposition, and logical form. New York:
Academic Press.
Giannakidou, Anastasia. 2005. N-words and negative concord. In M. Evaert, H. van Riemsdijk,
R. Goedemans, and B. Hollebrandse (Eds.), The Linguistic Companion, Volume 3,
(327-391). Oxford: Blackwell Publishing.
Giannakidou, Anastasia. 1999. Affective dependencies. Linguistics and Philosophy 22, 367-421.
Haegeman, Liliane. 1995. The Syntax of Negation. Cambridge: Cambridge University Press.
Haegeman, Liliane and Raffaella Zanuttini. 1996. Negative Concord in West Flemish. In
A. Belletti and L. Rizzi (Eds.) Parameters and Functional Heads. Essays in Comparative
Syntax. Oxford University Press: New York. p. 117-179.
Haiman, John 1991. From V/2 to Subject Clitics: Evidence from Northern Italian. In Elizabeth
Closs Traugott and Bernd Heine (Eds.) Approaches to grammaticalization: Focus on
29
theoretical and methodological issues. Amsterdam: John Benjamins Publishing
Company.
Hopper, Paul J. and Traugott, Elizabeth C. 1993. Grammaticalization. Cambridge: Cambridge
University Press.
Horn, Laurence R. 1985. Metalinguistic negation and pragmatic ambiguity. Language 61. 121174.
Horn, Laurence R. 1964. A Natural History of Negation. Chicago: Chicago University Press.
Jayaseelan, Karattuparambil. 2008. Topic, focus and adverb positions in clause structure.
Nanzan Linguistics 4: 43-68.
Jayaseelan, Karattuparambil. 2001. IP-internal topic and focus phrases. Studia Linguistica
55: 39-75.
Kandybowicz, Jason. 2013. Ways of emphatic scope-taking: from emphatic assertion in Nupe
to the grammar of emphasis. Lingua 128: 51-71.
Klima, Edward S. 1964. Negation in English. In Jerry A. Fodor and Jerrold J. Katz (Eds.) The
Structure of Language 246-323. Englewood Cliffs NJ: Prentice Hill.
Kroeger, Paul. 2014. External negation in Malay/Indonesian. Language 90:1, 137-184.
Laka, Itziar. 1994. On the syntax of negation. New York: Garland Publishing.
Laka, Itziar. 1990. Negation in syntax: On the nature of functional categories and projections.
Ph.D. thesis. Massachusetts Institute of Technology.
Lehmann, Christian 1982. Thoughts on Grammaticalization. Munchen: Lincoln.
Levinson, Stephen C. 1983. Pragmatics. Cambridge: Cambridge University Press.
McCawley, Jim. 1988. The Syntactic Phenomena of English. Chicago: Chicago University Press.
McClave, Evelyn Z. 2001. Spontaneous gestures of the hearing and American Sign Language.
Gesture 1.1, 51-72.
Neidle, Carol; Kegl, Judy; MacLaughlin, Dawn; Bahan, Benjamin; and Lee, Robert. 2000. The
Syntax of American Sign Language. Functional Categories and Hierarchical Structure.
Cambridge, MA: MIT Press.
Payne, John 1985. Negation. In T. Shopen (Ed.) Language Typology and Syntactic Description.
197-242. Cambridge: Cambridge University Press.
Pfau, Roland 2015. The grammaticalization of head shake: From head movement to negative
head. In Andrew D.M. Smith, Graeme Trousdale, and Richard Waltereit (Eds.) New
Directions in Grammaticalization Research 9-50. Amsterdam: John Benjamins
Publishing Company.
Partee, Barbara H. 2004a. Compositionality in Formal Semantics: Selected Papers of
Barbara Partee. Oxford: Blackwell Publishers.
Partee, Barbara H. 2004b. Applications of the lambda calculus to linguistic examples. Formal
Semantics Lecture Notes.
Poletto, Cecilia and Raffaella Zanuttini. 2013. Emphasis as reduplication: evidence from sì
che/no che sentences. Lingua 128: 124-141.
Roberts, Ian and Anna Roussou. 2003. Syntactic Changes: A Minimalist Approach to
Grammaticalization. Cambridge: Cambridge University Press.
Rooryck, Johan. 2008. A compositional analysis of French negation. Unpublished
manuscript, Leiden University.
Schlenker, Philippe. 2016. Super Monsters !!: Role Shift, Iconicity and Quotation in Sign
Language. Semantics & Pragmatics. To appear.
Stokoe, William. 1960. Sign Language Structure: An Outline of the Visual Communication
30
Systems of the American Deaf – Studies in Linguistics: O.P.8. (revised, 1978, Linstock
Press).
van der Auwera, Johan. 2006. Why languages prefer prohibitives. Wai Guo Yu [Journal of
Foreign Languages] 161: 1-25.
van der Auwera, Johan and Neuckermans, Annemie. 2004. Jespersen’s Cycle and the interaction
of predicate and quantifier negation. In Dialectology meets typology. Dialect grammar
from a cross-linguistic perspective. Bernd Kortmann (Ed.) 453-478. Berlin: Mouton de
Gruyter.
Veinberg, Silvana and Wilbur, Ronnie 1990. A linguistic analysis of the negative headshake in
American Sign Language. Sign Language Studies 68, 217-244.
Wilbur, Ronnie. 1999. Metrical Structure, Morphological Gaps, and Possible Grammaticalization
in ASL. Sign Language & Linguistics 2:2. 217-244.
Yoon, Suwon 2013. Parametric Variation in Subordinate Evaluative Negation: Japanese/Korean
vs. Others. Journal of East Asian Linguistics 22.2: 133-166.
Yoon, Suwon 2011. ‘Not’ in the Mood: The Syntax, Semantics, and Pragmatics of Evaluative
Negation. Ph.D. thesis. The University of Chicago.
Yoon, Suwon 2008. Syntax and Semantics of Negation. Current issues in Unity and Diversity of
Languages. Congress book of the 18th International Congress of Linguistics (CIL18),
2541-2558. The Linguistics Society of Korea, Seoul, Korea.
Zanuttini, Raffaella and Portner, Paul 2003. Exclamative Clauses: At the syntax-semantics
interface. Language 79(1), 39-81.
Zanuttini, Rafaella. 1991. Syntactic Properties of Sentential Negation: A comparative study of
Romance languages. Ph.D. dissertation, University of Pennsylvania.
Zeijlstra, Hedde. 2004. Sentential Negation and Negative Concord. Ph.D. thesis. University of
Amsterdam.
Zeshan, Ulrike 2004. Head, hand, and face: Negative constructions in sign languages. Linguistic
Typology 8, 1-58.
Zwarts, Frans. 1995. Nonveridical contexts. Linguistic Analysis 25, 3-4: 286-312.
31