Can't you see the difference? Sources of variation in sign language structure

Full text

Can’t you see the difference?
Sources of variation in sign language structure
Pamela Perniss, Roland Pfau, and Markus Steinbach
1. Introduction
Signed and spoken languages are produced and perceived in radically
different ways. While spoken languages are produced by the vocal tract and
perceived by the auditory channel, signed languages are produced by the
hands, but also other non-manual articulators like the head, face, and body,
and are perceived visually. Sign linguistic research in the past decades (see
Section 2 for a brief overview of the history of sign language research) has
proven beyond a doubt that natural language exists in two modalities, and
thus, that signed and spoken languages share basic linguistic properties on
the levels of phonological, morphological, and syntactic structure.
Still, modality plays an important part in shaping the expression of
linguistic structure. With respect to how modality can influence linguistic
structure, the role of iconicity or visual motivation is of particular
importance. The visual-gestural modality affords a much higher potential
for iconic representation than the auditory-vocal modality. The force of
iconicity is evident, for example, in indexical reference (see Cormier, this
volume), the use of space to represent location and motion of referents (see
Johnston et al., this volume), and referential shift (see Pyers and Senghas,
this volume). In addition to the role of iconicity, the nature of the visual-
gestural modality also affects other parts of linguistic structure. For
example, it provides the possibility of, and seems to favor, non-
concatenative morphology (Klima and Bellugi 1979, Aronoff et al. 2005).
Meier (2002) lists three other prominent differences between the two
language modalities that may cause differences in the linguistic structure of
signed and spoken languages: the different nature of the articulators used
for language production, the different nature of the perceptual systems used
for language comprehension, and the comparative youth of signed
languages. Thus, modality may affect linguistic structure, and indeed
properties of the visual-gestural modality have been argued to create a
homogenizing effect in sign languages, leading to less variation overall in

Pamela Perniss, Roland Pfau, and Markus Steinbach
2
sign language structure compared to the variation found across spoken
languages (Newport and Supalla 2000, Aronoff et al. 2005).
Until recently, research on sign languages was limited to American Sign
Language (ASL) and a number of European sign languages as, for example,
French, German, British, Swedish, and Danish Sign Language (cf. also
Section 2). The current research climate is testimony to a surge of interest
in the study of a geographically more diverse range of sign languages. This
volume reflects that climate and brings together work by scholars engaging
in comparative sign linguistics research. Before we can truly answer the
question of whether modality effects do indeed cause less structural
variation in sign languages as compared to spoken languages, it is
necessary to investigate the differences that exist between sign languages in
more detail and, especially, to include in this investigation less studied
(often non-Western) sign languages (see Zeshan 2004a, 2004b, 2006 for
pioneering work in this area).
In this spirit, the focus of the present volume is variation within the
modality of sign. The various contributions concentrate not on a specific
domain, but rather cover a range of different areas, including word pictures,
negation, auxiliaries, constituent order, sentence types, modal particles, and
role shift. One question that arises is whether the range and extent of
variation differs between linguistic domains, and, if yes, whether the
differences are attributable to properties of the modality. For example,
modality may affect some grammatical domains to a greater extent than
others. Likewise, the iconicity of signs and grammatical constructions may
decline over time, and different domains may be variously affected by such
processes.
Before turning to possible sources of variation at different linguistic
levels in Section 3, we will briefly sketch important developments in the
history of sign language linguistics in Section 2. Finally, Section 4 gives an
outline of the content of this volume.
2. Developments in sign language linguistics
In order to situate the discussion below as well as the contributions to the
present volume in a historical context, we will first say a few words about
important developments in sign language research. Obviously, the picture
sketched in this section is very much simplified. Still, we believe that the
research endeavours undertaken in the area of sign language linguistics

Sources of variation in sign language structure 3
since the 1960’s can roughly be divided into three periods characterized by
different theoretical objectives.1
In the first period of the study of signed language, researchers focused
on the underlying identity between spoken and signed languages. Woll
(2003) calls this period, which started in the middle of the twentieth
century, the “modern period”. Determined to prove the linguistic status of
sign languages against widely held prejudices and misconceptions that
communication between the deaf was based on pantomime and gesture,
early sign linguists de-emphasized the role of iconicity in sign language
(see, for instance, Klima and Bellugi 1979). This was the case for lexical
signs, but also notably for the system of classifiers. Studies have shown that
many lexical signs are characterized by an arbitrary form-meaning
mapping, and that the meanings of lexical signs cannot easily be guessed
by naïve non-signers (cf. Pizzuto and Volterra 2000). The predominant sign
language investigated in this period was ASL. As a consequence, there was
little typological research.
In the post-modern area starting in the 1980’s, researchers first turned to
the issue of modality and investigated similarities and differences between
signed and spoken languages. In this period, researchers were interested in
the influence of modality on linguistic structure, in modality-specific
properties of signed and spoken languages, and in modality-independent
linguistic universals. Starting from the observation that sign languages
seem to be typologically more homogenous than spoken languages, many
grammatical properties of sign languages have been related to specific
properties of the visual-gestural modality discussed in Section 1 above
(Meier 2002). In both the modern and the post-modern period, sign
language research mainly focused on the comparison of sign languages to
spoken languages. Cross-linguistic studies on sign languages have been
rare. However, the hypothesis that sign languages are typologically more
similar than spoken languages has to be taken with caution until more (non-
related) sign languages have been investigated (Woll 2003).
Only once non-Western sign languages entered the stage, it became
clear that sign languages show more variation than originally predicted.
This third period, which approached sign language typology more
seriously, started at the end of the 1990’s. Today, we can observe an
increasing interest in comparative studies on sign languages at all linguistic
levels that also include less studied (Western and non-Western) sign
languages. In this context, researchers also develop new methodological
and technological tools for the elicitation, collection, and documentation of

Pamela Perniss, Roland Pfau, and Markus Steinbach
4
sign language data (see Johnston et al., this volume). Still, more
comprehensive documentations and typological studies of different sign
languages are necessary for a better understanding of the similarities and
differences between sign languages in particular and signed and spoken
languages in general. In the long term, sign language typology is expected
to make an important contribution to a better understanding of the nature of
human language.
3. Sources of variation
Obviously, the research endeavors undertaken by the authors of this volume
belong to the third of the above-mentioned periods: the documentation of
similarities and differences between sign languages. In this section, we
briefly sketch a number of linguistic areas in which variation has been
found in order to give the reader a first impression of what forms sign
language variation may take. Many of the aspects tackled in this section
will be discussed in much more detail in contributions to this volume. The
list of topics presented in the following sections is by no means exhaustive.
However, we take the aspects we selected to be illustrative of the types of
variation found across sign languages. We shall look at three linguistic
levels of description in turn, considering first phonological (Section 3.1),
then morphological (Section 3.2), and finally syntactic variation (Section
3.3). More examples from these three domains as well as the issue of
lexical variation are discussed in Hohenberger (this volume).
3.1. Phonology
Since Stokoe’s (1960) seminal work on sign language structure, it is a well-
known fact that signs are not holistic units but are composed of smaller
phonological units often referred to as phonological parameters (‘cheremes’
in Stokoe’s terminology). While Stokoe himself identified three parameters
– handshape, location, and movement – later research proved the
importance of two further aspects, namely orientation and non-manuals.2 In
this section, we first discuss cross-linguistic variation in some of the
phonological parameters. We then turn to a phonological rule that has been
shown to be subject to language-specific constraints: weak hand drop (see

Sources of variation in sign language structure 5
Hohenberger, this volume, for discussion of variation in minimal syllable
sonority).
3.1.1. Phonological parameters
Clearly, the phonological building blocks of language are modality-
specific: consonants are simply not attested in sign languages and
handshapes do not play a role in spoken language phonology. Still,
researchers have shown that the internal and external organization of these
building blocks follows modality-independent principles; see, for example,
Sandler (1989) and Brentari (1998) for feature hierarchies and Perlmutter
(1992) for syllable structure.
Spoken languages vary considerably with respect to their phoneme
inventories. The question therefore arises: how much and what type of
variation exists in the phonological parameter inventories of sign
languages? In this section, we will briefly consider handshape, location,
movement, as well as non-manuals.3
The hand can be in various configurations, depending on whether and
how many fingers are selected, and on whether the selected fingers are
extended, bent, hooked, or curved. Different sign languages have different
inventories of handshapes. Variation in handshape inventories can be due to
two factors. First, while all known sign languages share a number of
handshapes – including at least the so-called ‘unmarked handshapes’ (cf.
Sutton-Spence and Woll 1999: 162) – there are some complex handshapes
that are only attested in few sign languages. Note that in this context, the
notion ‘complex’ refers to featural complexity, which is defined as the
number of distinctive features necessary to describe a handshape (cf.
Sandler 1996). The complex handshapes shown in Figure 1, for instance,
are infrequent.
Figure 1. Infrequent handshapes

Pamela Perniss, Roland Pfau, and Markus Steinbach
6
Secondly, sign languages vary in the size of their handshape inventories.
For example, compared to a sign language like ASL, Adamorobe Sign
Language (AdaSL), a village sign language in Ghana, has a very small
handshape inventory (Nyst 2007).
Signs can have fixed points of articulation on the face or body or can be
executed in neutral space, that is, in the area of space in front of the body.
The chest, the shoulders, the arm, the wrist, the neck, and different parts of
the head and face, including the ear, the mouth, the eye, the nose, the
forehead, the side of the head, and the top of the head are all places of
articulation for signs. Differences between sign languages in place of
articulation have been suggested by Klima and Bellugi (1979) in a
comparison of signs in Chinese Sign Language (CSL) and ASL.
Some signs involve movement of the hand and/or of the fingers. The
hand(s) can move in a straight or arc-shaped path and can be executed in
different directions such as sideways, forwards, or contralaterally across the
body. Local movements of the fingers can be, for instance, wiggling or
bending, opening or closing. Klima and Bellugi (1979) also give examples
of movement values, both movement of the hands and internal movement
of the fingers or wrist, that differ between Chinese and ASL.
Sign languages also differ in the size of signing space, that is, in the size
of the space in front of and around the body in which signs are executed.
Generally, signing space is taken to extend vertically from the top of the
head to the waist, and horizontally slightly past the shoulders on each side
and forward to about arm’s reach. Sign languages like AdaSL or Kata
Kolok, a village sign language in Bali, for example, have a much bigger
signing space than do Western Sign Languages. In these sign languages,
the arms extend maximally to all sides, including points behind the body.
This is probably related to the use of an absolute reference frame (co-opted
from the surrounding spoken language and gestural systems) and a focus on
the “here and now”. This variability in the size of sign space is different
from the expansion or restriction of sign space that is found in “shouting”
or “whispering” in sign language, respectively (Crasborn 2001; Liddell
2003; Uyechi 1996).
Finally, the use of phonological non-manual elements differs between
sign languages. These are typically mouthings derived from the
surrounding spoken language that accompany signs.4 The use of mouthings
in ASL, a sign language generally considered to make only little use of
phonological mouthings, is the subject of the investigation by Nadolske and
Rosenstock (this volume). In contrast to what has been claimed for ASL,

Sources of variation in sign language structure 7
German Sign Language (Deutsche Gebärdensprache, DGS) is known to
make frequent use of mouthings. In DGS, mouthings occur obligatorily, for
example, with nominal signs and can disambiguate between different
meanings of an identical sign (the DGS signs for PAINT, BUTTER, and
MARMELADE, for instance, differ only in the accompanying mouthed
element). DGS also uses mouthings to differentiate between types of things
for which the manual sign provides the basic level identification. Different
types of birds, for example, can be distinguished on the basis of the
mouthing alone, whereby the manual sign remains the same (Keller and
Rech 1993).
3.1.2. Constraints on two-handed signs and weak-hand drop
We now turn to two-handed lexical signs. It has been shown that two-
handed lexical signs are subject to two phonological well-formedness
conditions: the symmetry condition and the dominance condition (Battison
1974). The first condition specifies that when both hands move in a two-
handed sign – be it symmetrically or in alternation – they must have the
same handshape (balanced sign). Conversely, the second condition states
that when the two hands do not share the same specification for handshape
(unbalanced sign), then one of them must be stationary/passive and,
moreover, the specification of the passive hand is restricted to one of a
small set, the articulatorily simple, unmarked handshapes shown in Figure
2. These phonological constraints seems to be valid across sign languages,5
although they might not hold in the same way for some Southeast Asian
sign languages like, for example, Korean Sign Language (KSL) (Kang Suk
Byun, personal communication).
Figure 2. Frequent, unmarked handshapes
Sometimes, two-handed signs can be signed without the non-dominant (or
weak) hand; this type of phonological deletion process is referred to as

Pamela Perniss, Roland Pfau, and Markus Steinbach
8
‘weak drop’ (Padden and Perlmutter 1987). While this phenomenon is
attested across many sign languages, recent research has shown that the
types of signs that can undergo weak drop differ from sign language to sign
language. Comparing the weak drop patterns of ASL and Sign Language of
the Netherlands (Nederlandse Gebarentaal, NGT), Van der Kooij (2001)
finds that two phonological specifications that block weak drop in ASL,
namely [alternating movement] and [crossing] (that is, one or both hands
crossing the midsagittal plane), do not always block weak drop in NGT.
That is, the NGT signs in Figure 3, MATCH with alternating movement as
well as AUSTRIA, in which both hands cross the midsagittal plane, do both
allow weak drop. In contrast, in ASL, similar signs cannot be signed with
only the dominant hand (Battison 1974).6
MATCH AUSTRIA
Figure 3. NGT signs that allow weak drop
Moreover, and also in contrast to ASL, Van der Kooij reports that weak
drop in NGT is acceptable in most unbalanced signs. This discussion shows
that a phonological rule that appears to be part of the phonological system
of many sign languages may still be subject to language-specific conditions
of application.
3.2. Morphology
In sign languages, the phonological and the morphological component
closely interact, since virtually every phonological parameter can function
as a morpheme by itself. That is, morphological processes tend to involve
stem-internal changes rather than affixation. In the domain of inflection,
handshapes can function as classifier morphemes (Section 3.2.1),
movement alterations can express aspectual meaning, and with some verbs

Sources of variation in sign language structure 9
changes in orientation and/or direction of movement can indicate the
Source and Goal of the action expressed the verb (see Section 3.3.2 below).
Moreover, non-manual markers (e.g. puffed cheeks, pursed lips) are
capable of supplying adjectival or adverbial meaning. Besides these stem-
internal changes, reduplication has been shown to be a productive
morphological process in sign languages. Interestingly, in sign languages,
reduplication expresses the same meanings as it does in spoken languages
(Moravcsik 1978; Pfau and Steinbach 2006): aspectual modification (e.g.
habituality and iteration), plurality (see Section 3.2.2), and reciprocity (Pfau
and Steinbach 2005a). As far as derivation is concerned, for instance,
conversion processes have been described that only affect the movement
component (manner and frequency) of a stem (see Section 3.2.3). In
addition to pluralization, classification, and derivation, we will also
highlight some cross-linguistic differences in pronominalization (Section
3.2.4)
3.2.1. Classifiers
Classifier predicates are complex predicates that consist of handshape and
movement morphemes that combine in certain (morphosyntactically
constrained) ways to express information about the size and shape,
handling, location, and motion of referents. The handshape reflects salient
visual-geometric properties of a referent, and thereby ‘classifies’ the
referent with respect to inherent properties of size and shape or, in some
cases, semantic class. Two main types of sign language classifiers are entity
classifiers, where the hand represents a referent as a whole and encodes
salient features of the entity’s size or shape, and handling classifiers, where
the hand represents the handling or manipulation of a referent (e.g,
Engberg-Pedersen 1993; Emmorey, 2003).
The use of classifier predicates has been described for the majority of
sign languages studied so far (see Schembri (2003) for a comprehensive
overview). However, the existence of classifier predicates seems to hold
primarily for urban sign languages. AdaSL, for example, exhibits a limited
use of handling classifiers, and does not use entity classifiers, at all (Nyst
2007).
Though classifiers are used in similar ways in the sign languages in
which they exist, the specific classifiers themselves differ between sign
languages. The correspondences between classifier handshape and visual-

Pamela Perniss, Roland Pfau, and Markus Steinbach
10
geometric properties of the referent exist per convention, and thus vary
from sign language to sign language. For example, in DGS, a B-hand (see
Figure 4 below) held horizontally with the palm down is used to represent
the semantic class of four-wheeled vehicles such as cars, buses, and trucks;
two-wheeled vehicles such as bikes and motorcycles, on the other hand, are
represented with a vertically-held B-hand. In ASL, an even broader
semantic class of vehicles, including water vehicles, is represented with a
single handshape (see Figure 4). Finally, a third, altogether different
handshape is used in Jordanian Sign Language (Lughat il-Ishaara il-
Urdunia, LIU) for the semantic class of vehicles (Hendriks 2004).
DGS ASL LIU
Figure 4. Entity classifiers for vehicles
In general, there is more variation between entity classifiers across sign
languages, as they tend to be more arbitrary, and more strongly
conventionalized. Handling classifiers tend to be more iconic, representing
the relevant action (i.e. the handling of the relevant object) more directly.
Cross-linguistic evidence suggests that across sign languages, the
subsystem of entity classifiers is more strongly grammaticalized than that
of handling classifiers (see Zeshan 2003 for Indopakistani Sign Language,
IPSL).
Finally, some sign languages, especially Asian sign languages, have
classifiers that mark gender (see Fischer and Osugi 2000 on Japanese Sign
Language – Nihon Syuwa, NS). In gender classifier systems, a separate
handshape is used for male and female referents. In NS, like in other Asian
sign languages, an extended upright thumb is the classifier form used for
males, while an extended upright pinky is used for females (cf. also Section
3.2.4).
3.2.2. Pluralization of nouns
Browsing through some of the available grammatical descriptions of sign
languages, we find striking similarities when it comes to the pluralization

Sources of variation in sign language structure 11
of nouns. In most of the studies, reduplication is mentioned as a common
pluralization strategy. One possible exception in this respect is IPSL where
– according to Zeshan (2000) – only the sign CHILD is reduplicated with
some frequency, while for other nouns, no morphological distinction is
made between singular and plural forms.
In a typological study on pluralization, Pfau and Steinbach (2006) show
that while reduplication is indeed a common strategy in pluralization, it is
subject to a number of phonological constraints (see Hohenberger, this
volume, for details). The nature of these constraints, however, may differ
from sign language to sign language. In DGS, for instance, body-anchored
nouns cannot be reduplicated. That is, the plural form of a body-anchored
sign like GLASSES (Figure 5) is realized by zero marking and the plural
interpretation either has to be inferred from the context or has to be
expressed by a numeral or quantifier.
Figure 5. The DGS body-anchored noun GLASSES
It appears that in NGT and ASL, the application of plural reduplication is
less constrained. In both these sign languages, the sign GLASSES (which is
phonologically similar to the sign given in Figure 5) can be reduplicated.
While in NGT, this is done with only the dominant hand performing a short
repeated movement towards the body location, in ASL, the reduplication
can be performed with both hands moving in alternation.
In other words: a brief look at nominal plurals might lead us to conclude
that they are realized in a similar way across sign languages. Closer
inspection, however, reveals that while the basic means of realizing
plurality (reduplication and zero marking) may be the same, their
applicability is clearly subject to language-specific phonological
constraints.

Pamela Perniss, Roland Pfau, and Markus Steinbach
12
3.2.3. Derivation
While various inflectional processes in sign languages, such as aspectual,
number (see Section 3.2.1 above), and spatial inflection (see Section 3.3.2
below), are well-described, comparatively little is known about derivation
in sign languages. From the available research, it appears that derivational
processes – in particular, sequential ones – are scarce in general.
Aronoff et al. (2005) describe some sequential derivational processes in
ASL and Israeli Sign Language (ISL). For ASL, they report an agentive
suffix grammaticalized from the noun PERSON that may attach to various
verbs as, for example, in TEACH^AGENTIVE (‘teacher’). They point out that
although the suffixed forms may reduce to a single movement contour
(which corresponds to one syllable), “the hand configuration and place of
articulation of each of the two morphemes are usually retained” (Aronoff et
al. 2005: 312).7 In ISL, they discovered a set of ‘sense prefixes’ which
consist of pointing to a sense organ (or the head or mouth). Many of the
resulting prefixed forms can be glossed as ‘to X by seeing (eye)/hearing
(ear)/thinking (head)/intuiting (nose)/saying (mouth)’. An example given
by the authors is the combined form EYE^SHARP meaning ‘to discern
visually’. This derivational process appears to be unique to ISL.
For both ASL and ISL, Aronoff et al. (2005) describe a negative suffix.
Form and use of the two suffixes, however, differ between the two sign
languages. The ASL suffix ZERO probably originates from the
phonologically similar sign NOTHING; it is signed with one hand in which
the fingers form the shape of a zero and it usually attaches to verbs
(SEE^ZERO ‘not see at all’). In contrast, the ISL suffix NOT-EXIST attaches
to adjectives (INTERESTING^NOT-EXIST ‘of no interest’) and has two
allomorphs – a one-handed and a two-handed one – the choice of which
depends on the form of the base sign (see Hendriks, this volume, for
discussion of a similar suffix in LIU).
From this brief discussion, we can conclude that some variation is
attested in the few sequential derivational processes described to date. The
same holds for simultaneous processes. While diminutive formation by
means of non-manual marking (pursed, rounded lips), for instance, is
probably found in all sign languages, other processes appear to be sign
language-specific. A case in point are the ASL ‘characteristic adjectival
rule’ and the ‘ISH adjective rule’ described in Padden and Perlmutter
(1987) both of which involve a change in movement pattern such as
repetition of movement and/or tense movement.

Sources of variation in sign language structure 13
Supalla and Newport (1978) found that in ASL, a change in movement
pattern also characterizes a fair amount of noun-verb pairs. In particular,
they show that verbs can have simple or repeated movement and moreover,
the movement may either end in a hold or be continuous. The noun-verb
pair SIT is an example for the former, while FLY is an example for the latter.
In the corresponding nouns, however, movement is repeated and tense
(‘restrained’ in their terminology), as can be seen in the noun signs CHAIR
and PLANE in Figure 6.
SIT CHAIR FLY PLANE
Figure 6. Verb-noun pairs in ASL
Recent research into noun-verb pairs in NGT has shown that in NGT the
patterns are not as clear as in ASL (Schreurs 2006). Many verbs and
corresponding nouns appear to be identical in form. Interestingly, for the
few standardized signs for which a systematic difference was found (for
example CIGARETTE/SMOKE and PLANE/FLY), the pattern is exactly the
opposite of the one described for ASL: the movement of the verb is tense
and repeated while the noun has continuous movement.8
3.2.4. Pronominal systems
As opposed to pronominal systems in spoken languages, pronominal
systems in sign languages seem to be quite uniform (McBurney 2002). The
pronominal systems of sign languages are determined to a large degree by
iconicity in the sense of indexicality, or actual pointing to their referents. In
the case of physically present referents, pronominal or indexical signs do
literally point to their referents, e.g. the signer points to her/his own chest to
indicate “I” and points to her/his interlocutor’s chest to indicate “you”, and
can likewise point to other animate or inanimate referents in the physical
context of the utterance. Non-present discourse referents can be
pronominally referred to by associating them with, and then pointing to,
particular locations in sign space.

Pamela Perniss, Roland Pfau, and Markus Steinbach
14
In Western sign languages, singular pronominal reference seems to be
made with an index finger point. These sign languages do not mark gender
on pronouns. By contrast, gender distinctions can be found in the
pronominal system of Asian sign languages, which incorporate gender
classifiers to distinguish between female and male pronouns (cf. McBurney
2002 and Section 3.2.1 above). In addition, the paradigms of plural
pronouns seem to show variation across sign languages with respect to the
degree of indexicality, the number and type of plural pronouns that exist,
and the types of plural inflection, i.e. movement modifications such as a
sweeping arc, that exist (see the comparison of first person plural pronouns
in ASL and British Sign Language (BSL) by Cormier (this volume)).
In addition to variation in the systems of personal pronouns, sign
languages also appear to exhibit considerable variation in their paradigms
of possessive pronouns. Again, variation exists in the number and type of
possessive pronouns that exist, in their syntactic distribution, as well as in
marking such distinctions as alienable vs. inalienable (cf. Neidle et al. 2000
and Sutton-Spence and Woll 1999).
3.3. Syntax
Not surprisingly, variation amongst sign languages is most striking when
we enter the realm of syntax. After all, the merging of a syntactic phrase
structure is highly abstract and independent of phonological properties of
the items to be inserted – no matter whether your theory involves
movement operations or not. Still, in this area, too, there are intriguing
similarities such as, for instance, the use of space for establishing syntactic
relations and the use of non-manual markers to distinguish sentence types.
In this section, we will discuss variation in constituent order (Section
3.3.1), in the use of agreement auxiliaries (Section 3.3.2), in the expression
of sentential negation (Section 3.3.3), in the realization of questions
(Section 3.3.4) and relative clauses (Section 3.3.5), and in the use of
signing space (Section 3.3.6).
3.3.1. Constituent order
It is a well-known fact that many of the sign languages investigated so far
allow for a fairly flexible constituent order. This has led some researchers

Sources of variation in sign language structure 15
to claim that constituent order in sign languages is relatively free (see
Friedman 1976 for ASL) or even that sign languages in general are not
characterized by an underlying hierarchical phrase structure (Bouchard and
Dubuisson 1995).
Others, however, have argued that once the existence of clause-external
material, such as topics and right-dislocated pronominals, and null
arguments is taken into consideration, it is very well possible to identify an
underlying, unmarked sign order. Consider, for instance, the examples in
(1). In the ASL example (1a), the object has been topicalized (as indicated
by the non-manual marker) and the resulting sign order is OSV (Neidle et
al. 2000: 50). In the NGT example in (1b), the surface sign order is OVS;
this order, however, is due to pronominal right dislocation of the subject
pronoun accompanied by pro drop. Crucially, full arguments cannot appear
in post-verbal position.
top
(1) a. JOHNi, MARY LOVE ti [ASL]
‘John, Mary loves.’
b. pro BOOK BUY INDEX3a [NGT]
‘He buys a book.’
Other factors that have been shown to have an impact on the order of signs
in a sentence are the semantic reversibility of arguments (Coerts 1994) and
morphosyntactic characteristics of the verb, such as aspectual and spatial
inflections labelled “reordering morphology” by Chen Pichler (2001).
Once the influence of these factors is acknowledged, it turns out that
ASL has an underlying SVO-order while the basic order in NGT is SOV.
That is, sign languages may obviously differ from each other with respect
to constituent order. Other sign languages that are claimed to display SVO-
order include Brazilian Sign Language (Língua de Sinais Brasileira, LSB),
Hong Kong Sign Language (HKSL), and Swedish Sign Language (SSL);
other sign languages of the SOV-type are DGS, IPSL, and Italian Sign
Language (Lingua Italiana dei Segni, LIS) (see Johnston et al., this volume,
for discussion of constituent order in Australian Sign Language, Flemish
Sign Language, and Irish Sign Language; see Hohenberger, this volume,
for comparison of ASL and LSB). Note that so far no sign language with an
underlying VSO-order has been found – in contrast to spoken languages
where this order is not uncommon (Tagalog and Irish are two examples for
VSO-languages).9

Pamela Perniss, Roland Pfau, and Markus Steinbach
16
Moreover, even within the SVO- and SOV-group, sign languages may
differ from each other with respect to constituent order at the clause level.
Two sign languages that are both SOV, for instance, may display
differences in the positioning of modals (second position vs. post-verbal),
negative particles (see Section 3.3.3), or wh-signs (see Section 3.3.4).10
3.3.2. Agreement auxiliaries
Virtually all sign languages studied so far make a basic distinction between
agreement verbs (also called directing or indicating verbs) and plain verbs
(Padden 1988).11 Verbs of the first type can change phonological properties
(orientation and/or direction of movement) in order to signal which
participant is subject and object of the sentence (or, in terms of thematic
roles, Source and Goal of the action described by the verb). This option is
not available for verbs of the second type which are incapable of adapting
their form to the location of participants in that way.
In many sign languages, constituent order can be indicative of what
argument is the subject or object of the clause in case the clause contains a
plain verb. Some sign languages, however, have developed an alternative
strategy for indicating the grammatical role of arguments: they make use of
an auxiliary-like element that expresses the grammatical relations whenever
the lexical predicate is not capable of doing so. Consider the two examples
in (2) for illustration. The Taiwan Sign Language (TSL) verb LIKE is a
plain verb; in (2a), the auxiliary AUX2 moves in space from the locus of the
subject WOMAN towards the signer (Smith 1990: 220). Similarly, in the
DGS example (2b), the auxiliary glossed as PAM (person agreement
marker) accompanies the adjectival predicate ANGRY, thereby showing who
is angry with whom.
(2) a.
THAT FEMALE 3AUX21 LIKE [TSL]
‘That woman likes me.’
y/n
b. YESTERDAY INDEX2 TEACHER INDEX3b ANGRY 2PAM3b [DGS]
‘Were you angry with the teacher yesterday?’
Other sign languages that make use of similar auxiliary elements include
Catalan Sign Language (Llengua de Signes Catalana, LSC), Argentine
Sign Language (Lengua de Señas Argentina, LSA), and Greek Sign

Sources of variation in sign language structure 17
Language (GSL), while ASL, HKSL, and BSL are examples of sign
languages that do not have such an element available to them (see
Steinbach and Pfau, this volume, for details on the form, use, and
grammaticalization of agreement auxiliaries across sign languages).
3.3.3. Negation
As is true for other properties discussed in previous sections, the
similarities amongst sign languages are quite conspicuous when it comes to
the expression of sentential negation. A characteristic that has been noted
repeatedly in the literature is the combination of a manual negation sign
with a non-manual marker, viz. a side-to-side headshake. Based on this
observation, some researchers have argued that from a typological point of
view, these sign languages exhibit split negation where one element is a
particle and the other one a non-manual affix (Pfau 2002; Pfau and Quer,
this volume).
More recently, some interesting differences between sign languages
have been noted (Pfau and Quer 2002; Zeshan 2004a). On the one hand, the
position of the manual negative sign in the clause may vary from sign
language to sign language. It appears that, to some extent, the position of
this element is influenced by the basic sign order: in SOV languages, there
is a strong tendency for the manual negator to occupy the post verbal
position.12 On the other hand, and this is the more intriguing observation,
sign languages may also differ from each other with respect to the co-
occurrence of the manual and the non-manual element. Two aspects are
relevant here; since both of these are addressed in more detail in papers in
this volume, we will only mention them briefly.
First, the exact position of the headshake, its spreading characteristics, is
subject to different constraints across sign languages. For instance, while in
some sign languages, it is possible to have headshake on the manual
negative sign only, as illustrated in the HKSL example in (3a), in others the
headshake must at least extend over the predicate (for example, DGS; see
Pfau and Quer, this volume). Secondly, while in many sign languages, it is
possible, and actually quite common, to drop the manual sign and to negate
a proposition by means of a headshake only, in other sign languages, the
reverse pattern is observed: the manual negator is obligatory while the
headshake is optional. HKSL, LIS, and Turkish Sign Language (Türk İşaret
Dili, TİD), for instance, have been claimed to make use of such “manual-

Pamela Perniss, Roland Pfau, and Markus Steinbach
18
dominant” (Zeshan 2006) systems. For that reason, the HKSL utterance in
(3b) with non-manual negation only is ungrammatical (Tang 2006: 217;
also see Hendriks, this volume).
hs
(3) a. INDEX3 TOMORROW FLY NOT [HKSL]
‘It is not true that he is flying tomorrow.’
hs
b. * YESTERDAY NIGHT FATHER FAX FRIEND
‘Father didn’t fax his friend last night.’
Note finally that, while the use of a negative headshake – be it obligatory or
optional – has been attested in all sign languages investigated so far, some
sign languages also make use of backward head tilts to signal negation
(Zeshan 2004a; Hendriks, this volume). Clearly, we are dealing with the
grammaticalization of a culture-specific gesture here.
3.3.4. Question formation
Just as sentential negation discussed in the previous section, questions also
combine manual and non-manual marking (Petronio and Lillo-Martin 1997;
Neidle et al. 2000). Again, manual marking seems to show more variation
than non-manual marking. This is confirmed by Zeshan’s (2004b)
extensive cross-linguistic study on question formation in thirty-five sign
languages. While the use of non-manual markers in questions is very
similar across all sign languages investigated in this paper, the use of
manual markers (question particles), the structure of question-word
paradigms, and word order in interrogatives show more variation.
Let us turn to non-manuals in interrogatives first. Sign languages use
various non-manual means to indicate interrogatives, for instance eyebrow
position, eye contact with the addressee, and change in head and body
posture. Although all sign languages seem to use non-manuals to indicate
polar and wh-question, we also find some variation in this area. First,
different sign languages may use different kinds of non-manuals in
questions (see, for example, Šarac et al., this volume). Second, in many
sign languages, the non-manuals used in polar questions differ from the
non-manuals used in wh-questions. DGS, for example, uses raised
eyebrows for polar questions and lowered eyebrows for content or wh-

Sources of variation in sign language structure 19
questions. However, some sign languages, as for example HKSL, use the
same facial expression for both kinds of questions (Zeshan 2004b: 22).
Third, sign languages may differ in the scope of non-manuals. Both
examples in (4) are wh-questions without a wh-expression. Similar
examples can be found in many sign languages. In the NGT example in
(4a), the non-manual marker takes scope over the whole clause (Coerts
1992). By contrast, the NS example in (4b) shows that NS uses a specific
non-manual marker in clause-final position (Fischer and Osugi 1998).
wh
(4) b. MY SUITCASE [NGT]
‘Where’s my suitcase?’
wh
a. COLOR LIKE [NS]
‘What color do you like?’
Note finally that variation also results from the fact that some sign
languages do not only use non-manual means but also manual question
particles, while others have only non-manual question means at their
disposal. Zeshan’s study shows that between a fourth and a third of all sign
languages use question particles.
Question particles lead us to the issue of manual question markers in
sign languages. In a number of sign languages, a palm-up gesture is used as
a question particle. However, some sign languages have developed other
kinds of question particles. Spanish Sign Language (Lengua de Senãs
Espanõla, LSE), for example, uses the question particle SI/NO, which is
performed with an extended index finger signing first SI and then NO. Some
sign languages have even more than one question particle. HKSL, for
instance, distinguishes between the existential question particle HAVE-NOT-
HAVE and its non-existential counterpart GOOD-NOT-GOOD. While most
sign languages that have question particles use them only in polar
questions, some sign languages, like NGT, use them also in wh-questions.
The NGT question particle PALM-UP optionally appears in sentence-final
position in yes/no-questions (5a) and wh-questions (5b) (Coerts 1992;
Aboh and Pfau, in press).
y/n
(5) a. INDEX3 PARTY CANCEL INDEX3 PALM-UP [NGT]
‘Is the party cancelled?’

Pamela Perniss, Roland Pfau, and Markus Steinbach
20
wh
b. MARKET BUY WHAT PALM-UP [NGT]
‘What did you buy at the market?’
According to Zeshan (2004b), cross-linguistically the preferred position for
this particle is the clause-final position, but in some sign languages, it may
also appear sentence-initially or in both these positions.
A similar range of variation can be found in the syntactic distribution of
wh-expressions. In most sign languages, wh-words can appear in clause-
initial position, in clause-final position or in both positions simultaneously
(see also Šarac et al., this volume). By contrast, in IPSL, the placement of
the general question word is much more restricted. The general wh-sign
G-WH only occurs in sentence-final position (cf. Aboh et al. 2005).
Wh-word paradigms are another source of variation ranging from very
simple paradigms to highly complex ones. Interestingly, even sign
languages with complex wh-word paradigms usually have a general wh-
sign basically meaning ‘what’. Zeshan (2004b) therefore distinguishes three
different types of languages: (i) the general interrogative covers the whole
wh-word paradigm (type A), (ii) the general interrogative covers part of the
wh-word paradigm (type B), and (iii) the general interrogative exists
alongside a complex wh-word paradigm (type C). IPSL belongs to type A
since it has only the general wh-sign G-WH, which can be combined with
non-interrogative signs to derive more specific complex wh-expressions
such as, for example, FACE + G-WH meaning ‘who’. LSB is a type B
language with three specific wh-signs (‘how’, ‘why’, and ‘how many’).
Finally, type C languages with complex wh-word paradigms are, for
example, ASL and DGS.
3.3.5. Relative clauses
In spoken languages, relative clause constructions are known to show
considerable variation (Keenan 1985; Lehmann 1986). Among others, the
following parameters distinguish relative clauses across languages: (i)
position of head: externally vs. internally headed relatives, (ii) type of
relative construction: relative clauses vs. correlatives, and (iii) the use of
specific markers: relative pronouns, relative complementizers, or
resumptive pronouns.

Sources of variation in sign language structure 21
Although so far, relative clauses have only been investigated in detail
for three sign languages, ASL, LIS, and DGS, the same range of variation
has been found as in spoken languages. While in all three sign languages, a
non-manual marker (raised eyebrows) is used to indicate relative
constructions, the syntactic properties of relative constructions differ from
sign language to sign language. Head-internal relative clauses, for example,
are attested in ASL. In (6a) the head noun DOG is clearly part of the relative
clause, as evidenced by the fact that the adverbial precedes the head noun
and the non-manual marker extends over the head noun (Liddell 1978).
Note that the sentence is ambiguous: while it is clear that the dog chased
the cat, it is not clear which of the two animals came home. DGS, on the
other hand, uses head-external relative clauses, as illustrated in example
(6b), in which the head noun WOMAN appears outside the relative clause.
The relative clause itself is introduced by the relative pronoun RPRO-H and
the non-manual extends only over the relative pronoun (Pfau and Steinbach
2005b).13
rel
(6) a. [RECENTLY DOG (THATa) CHASE+ CAT] COME HOME [ASL]
‘The dog which recently chased the cat came home.’
‘The cat which the dog recently chased came home.’
rel
b. WOMAN [RPRO-H3a MAN IX3b 3aHELP3b] KNOW 3aPAM1 [DGS]
‘The woman who is helping the man knows me.’
c. [YESTERDAY HOUSEi MARIA SEE PRORELi] TODAY BURN [LIS]
‘The house Maria saw yesterday burnt today.’
Yet another type of relative construction has been described for LIS.
Cecchetto et al. (2006) analyze LIS relative constructions such as (6c) as
head-internal correlative constructions containing the clause-final
correlative marker PROREL.14 According to these authors, the extension of
the non-manual marker (not given for (6c)) is variable.
The above examples also exemplify another domain of variation in sign
language relative clauses: the use of manual relative markers. Sign
languages, like spoken languages, may use relative complementizers,
relative pronouns, and zero marking. According to Liddell (1978), relative
complementizers are attested in certain relative clauses in ASL (the
optional marker THATa in (6a)). Relative pronouns and a correlative marker

Pamela Perniss, Roland Pfau, and Markus Steinbach
22
are used in DGS and LIS, respectively, whereas relative clauses without a
manual marker are found in LSB and in ASL.
3.3.6. The use of signing space
As already noted in section 3.2.1 above, the location, orientation, and
motion of classifier predicates in sign space can indicate the location,
orientation, and motion of objects in the real world. That is, the locations of
classifiers in sign space schematically correspond to the locations of objects
in the environment or event space being described. This topographic use of
sign space is one of the most unique features of the visual-gestural
modality, and is taken to be a general affordance of this modality.
In addition to the use of classifier forms, the way spatial relationships
are represented in sign space is dependent on the viewpoint or perspective
the signer takes. On the one hand, signers can assume a global viewpoint
and oversee the entire environment or event space from an external
perspective. On the other hand, the signer can take an event-internal
perspective by assuming the role of a participant within the event (as in role
shift or constructed action, cf. Liddell and Metzger 1998). These two types
of mapping have been described by numerous researchers using different
terminologies: Liddell (2003) distinguishes between “depictive space” and
“surrogate space”; Morgan (1999) uses the terms “fixed referential
framework” and “shifted referential framework”; Schick (1990) describes
the use of “model space” and “real-world space”; Emmorey and Falgier
(1999) distinguish the use of “diagrammatic space” and “viewer space”;
and Perniss and Özyürek (in press) use the terms “observer perspective”
and “character perspective”, respectively.
The use of these devices, especially the use of classifier predicates, has
been assumed to be similar across sign languages due to the assumption of
modality effects driven by the iconic properties of sign languages (Meier
2002; Talmy 2003; Aronoff et al. 2005). However, there has been little
research on the way referent location, motion, and action is represented in
sign space using classifier predicates, as well as other spatially modifiable
signs like index signs and indicating verbs.
In a preliminary study comparing the use of classifier predicates and
perspective in event representations in DGS and TİD), Perniss and Özyürek
(in press) show that these two sign languages appear to impose different
linguistic or discourse constraints on the use of space to depict referent

Sources of variation in sign language structure 23
location, motion, and action. For example, contrary to what was observed
for TİD signers, DGS signers seem to disprefer the use of handling
classifiers in a spatial representation from an observer’s perspective.
Overall, the results indicate that this domain, where modality effects are
widely considered to create similarities in the use of space across sign
languages, may exhibit more variation than previously thought. The results
of the study comparing referential shift marking in ASL and Nicaraguan
Sign Language (NSL) presented by Pyers and Senghas (this volume)
likewise suggest that sign languages can conventionalize a range of
different devices and use space in various ways within this system.
4. Content of this book
The articles in this volume take up many of the topics discussed in the
previous sections and also add new topics. They discuss data from many
different sign languages (for an overview see section 2 of the notational
conventions) and cover a wide range of topics from different areas of
grammar including phonology (word pictures), morphology (pronouns,
negation, and auxiliaries), syntax (word order, interrogative clauses,
auxiliaries, negation, and referential shift) and pragmatics (modal meaning
and referential shift). In addition to this, one paper addresses
psycholinguistic issues (slips of the hand) and three papers deal with
aspects of language change (grammaticalization). In addition to this, many
papers discuss issues concerning data collection in sign languages and
provide methodological guidelines for further research. Although some
papers use a specific theoretical framework for analyzing the data, this
volume clearly focuses on empirical and descriptive aspects of sign
language variation.
The paper by Marie A. Nadolske and Rachel Rosenstock is the only one
in the volume that looks at, or rather reconsiders, phonological variation. In
their study, the authors investigate the occurrence of mouthings in ASL.
Mouthings are mouth movements which resemble spoken words and
accompany manual signs. In the past, it has been claimed that ASL uses
mouthings to a much lesser degree than European sign languages. Nadolske
and Rosenstock, however, provide evidence that mouthings are frequently
used in ASL across various discourse situations. Additionally, they show a
relationship between the occurrence of mouthings and word classes.

Pamela Perniss, Roland Pfau, and Markus Steinbach
24
In her investigation of pronoun indexicality, Kearsy Cormier explores a
domain in which the potential of the visual-gestural modality for iconic
representation plays a strong role. The article compares first person plural
pronouns in ASL and BSL and investigates the extent to which these
pronouns actually index (point toward) the locations associated with their
referents. Cormier looks at both inclusive and exclusive contexts and shows
that first person plural pronouns in the two sign languages exhibit variation
with respect to indexicality. She discusses the loss of indexicality in
exclusive pronouns, in particular, and offers explanations based on both
linguistic and motor factors. The paper is an important contribution to our
understanding of the ways in which the form of iconic or highly visually
motivated signs can be constrained within a conventionalized linguistic
system.
Bernadet Hendriks’ contribution adds to our understanding of the
variation in the expression of sentential negation by discussing data from
an as yet under-investigated sign language, namely Jordanian Sign
Language (Lughat il-Ishaara il-Urdunia, LIU). She reports on the
distribution of various manual negative signs (including negative concord),
on morphological negation by means of a suffix, and on the use of non-
manual markers in negation. A comparison to negative structures in other
sign languages (ASL, CSL, DGS, and LSC) reveals interesting cross-
linguistic differences with respect to the obligatory presence of a manual
negator, the nature and use of non-manual markers, and the possibility of
negative concord.
The second paper dealing with negative structures is the one by Roland
Pfau and Josep Quer. They add to the findings of an earlier comparative
study on sentential negation in DGS and LSC by reporting on the use and
distribution of negative modals in the two sign languages. It turns out that
while DGS and LSC – both SOV-languages – show fine-grained
differences in the distribution of the negative headshake in clauses with
lexical predicates, they pattern alike in negative clauses containing modals.
Pfau and Quer propose a generative grammar analysis to account for the
observed similarities and differences.
Trevor Johnston, Myriam Vermeerbergen, Adam Schembri, and
Lorraine Leeson present a cross-linguistic study of constituent ordering in
Flemish Sign Language (VGT), Irish Sign Language (ISL), and Australian
Sign Language (Auslan). In addition to providing valuable data about sign
language variation in this central syntactic domain, their paper discusses
important issues concerning data collection and analysis. Based on an

Sources of variation in sign language structure 25
overview of previous studies on constituent order and their own small-scale
cross-linguistic study, the authors point out difficulties for cross-linguistic
comparisons due to different methodology and terminology, even when the
same elicitation materials are used. Their own comparison is dedicated to
ensuring comparability and accessibility of language data, and provides
clear methodological guidelines.
In contrast to most areas of sign language linguistics, the syntax of
questions is a field that is comparably well studied from a theoretical and
typological point of view (cf. section 3.3.4). Still, more sign languages need
to be investigated to yield a more fine-grained picture of possible
interrogative constructions in sign languages. In their paper, Ninoslava
Šarac, Katharina Schalber, Tamara Alibašić, and Ronnie B. Wilbur focus
on interrogatives in two less studied European sign languages, Croatian
Sign Language (Hrvatski Znakovni Jezik, HZJ) and Austrian Sign
Language (Österreichische Gebärdensprache, ÖGS), and compare them to
interrogatives in ASL. The paper addresses manual and non-manual
interrogative markers. In all three sign languages, polar and wh-questions
are marked non-manually and different markers for polar and wh-questions
are used. Moreover, the wh-sign can occur in sentence initial, sentence
final, or in both positions. Interestingly, HZJ and ÖGS use the same non-
manual marker, which differs from the marker used in ASL, whereas only
ASL and HZJ have an additional manual marker for polar question at their
disposal.
In her paper, Annika Herrmann breaks new ground by considering
variation within the expression of pragmatic aspects of utterances. She
discusses the expression of the speaker’s attitude towards the utterance
(which is often called modal meaning) in two spoken (English and German)
and two signed languages (DGS and Irish Sign Language, ISL).
Herrmann’s study reveals that the two sign languages show less variation in
the expression of modal meaning than the two spoken languages.
Nevertheless, it also turns out that the extent of variation between the two
sign languages is greater than expected. Whereas in both sign languages,
non-manual features are the basic means of indicating the speaker’s
attitude, ISL also uses various manual and gestural expressions to mark
modal meanings. Moreover, Herrmann shows that the non-manual features
used in ISL differ from the ones used in DGS.
The contribution by Jennie E. Pyers and Ann Senghas compares the
system of referential shift in ASL, a well-established sign language, and
Nicaraguan Sign Language (NSL), a young, emerging sign language. The

Pamela Perniss, Roland Pfau, and Markus Steinbach
26
authors show that there are differences between the two sign languages in
the devices used to mark referential shift, and in the maintenance of
discourse cohesion through spatial mapping. The differences found
between ASL and NSL are discussed in light of the relative youth of NSL,
as the differences in the use of devices by NSL signers of different ages
suggest that this young sign language is in the process of developing a
more strongly conventionalized means of marking referential shift. In
addition, the authors address the possible influence of the gestural systems
of the surrounding spoken languages on the development of the ASL and
NSL systems of referential shift.
Markus Steinbach and Roland Pfau investigate the diachronic
development of a sign language-specific kind of auxiliary, so-called
agreement auxiliaries. As opposed to common auxiliaries found in spoken
languages, agreement auxiliaries do not encode tense, aspect, or modality
but subject and object agreement (cf. section 3.2.2 above). The authors
show that (i) agreement auxiliaries are attested in many (unrelated) sign
languages and (ii) that sign languages use modality-specific
grammaticalization paths for the development of auxiliaries. In sign
languages, unlike in spoken languages, auxiliaries develop not only from
verbal sources but also from nominal and pronominal ones. Steinbach and
Pfau argue that this difference between spoken and signed languages results
from spatial (phonological) and certain semantic properties of agreement in
sign languages. Pronouns and certain nouns provide optimal sources for the
grammaticalization of agreement auxiliaries.
In the final paper of this volume, Annette Hohenberger addresses the
issue of possible variation between sign languages from a more theoretical
point of view. Before turning to attested variation in several linguistic
domains (phonology, morphology, syntax, and lexicon), she discusses
possible determinants of linguistic variation in general: (i) general cognitive
properties of representation and processing, (ii) general task properties, (iii)
principles and parameters of Universal Grammar, (iv) typology, and (v)
modality. She adds to the picture the results of research into sign language
processing, that is, slip of the hand data from DGS and ASL. She suggests
to draw on a comprehensive theory of the human language faculty such as
generative grammar which claims universal representations and processes
that allow for an abstract model-theoretic characterization of the structure
and the processing of a language.

Sources of variation in sign language structure 27
Notes
1. This rough division of research is, of course, not meant to imply that all
studies on sign language in one period follow the respective predominant
paradigm. Also note that we confine ourselves to core linguistic aspects only.
We will not consider psycho- and neurolinguistic as well as social and
institutional issues (for a more detailed discussion of the history of sign
language linguistics, see Woll 2003).
2. In some models, handshape (selected fingers and position of fingers) and
handorientation are subsumed under a handconfiguration node (see, for
instance, Sandler 1989 for ASL).
3. At present, we are not aware of variation that would concern orientation (of
the fingers and palm).
4. For variation in other kinds of non-manuals see section 3.3 below.
5. See, for instance, Pfau (1997) for DGS, van der Kooij (2001) for Sign
Language of the Netherlands (Nederlandse Gebarentaal, NGT), and Sutton-
Spence and Woll (1999) for British Sign Language (BSL).
6. The sign
MATCH is taken from www.gebarencentrum.nl, the sign AUSTRIA
from www.effathaguyot.nl. Note that the ASL sign AUSTRIA is identical to the
NGT sign given in Figure 3.
7. A similar element is attested in German Sign Language (Deutsche
Gebärdensprache, DGS) and NGT; still, for these two sign languages, it is not
clear at present whether the morphological process is one of derivation or
compounding.
8. Schreurs (2006) also found a difference in the non-manual component of
standardized NGT nouns and verbs: while almost all nouns are accompanied
by a mouthing (i.e. a silent articulation of (part of) a Dutch word), almost all
of the verbs are accompanied by a mouth gesture (i.e. a mouth movement that
is not related to the spoken language). See Nadolske and Rosenstock, this
volume, for further discussion of mouthing.
9. The fact that no known sign language exhibits an underlying order in which
the object would precede the subject (VOS, OVS, or OSV) is less surprising
since these orders are also very rare across spoken languages.
10. Sign languages also differ from each other with respect to the sign order in the
nominal domain, that is the position of determiners, adjectives, numerals, and
quantifiers vis-à-vis the head noun. We will not go into this issue here.
11. Kata Kolok, a village-based sign language of Bali, seems to be an exception to
this generalization. Marsaja and Kanta (2005) point out that the only verbs in
the sign language that are used directionally with some frequency are the
verbs GIVE and TAKE.

Pamela Perniss, Roland Pfau, and Markus Steinbach
28
12. As pointed out by Zeshan (2004a), sign languages also differ with respect to
the size of their paradigm of clause negators. While all sign languages appear
to have a negative particle that conveys basic clause negation, some have at
their disposal other manual negators with a more specialized meaning, such as
negative existentials, negative modals, negative completives, or negative
imperatives.
13. Note that DGS has two relative pronouns: RPRO-H is used for human referents
and RPRO-NH for non-human referents.
14. But see Branchini and Donati (in press) whose analysis of relative
constructions in LIS slightly differs from the analysis proposed in Cecchetto
et al. (2006). Branchini and Donati argue that LIS relative constructions are
best analyzed as internally headed relative clauses, although they share many
properties with correlatives.
References
Aboh, Enoch and Roland Pfau
in press What’s a wh-word got to do with it? In The cartography of CP,
Paola Benincà and Nicola Munaro (eds.). Oxford: Oxford University
Press.
Aboh, Enoch, Roland Pfau, and Ulrike Zeshan
2005 When a wh-word is not a wh-word: The case of Indian Sign
Language. In The Yearbook of South Asian languages and linguistics
2005, Tanmoy Bhattacharya (ed.), 11-43. Berlin: Mouton de
Gruyter.
Aronoff, Mark, Irit Meir, and Wendy Sandler
2005 The paradox of sign language morphology. Language 81: 301-344.
Battison, Robbin
1974 Phonological deletion in American Sign Language. Sign Language
Studies 5: 5-19.
Bouchard, Denis and Colette Dubuisson
1995 Grammar, order and position of wh-signs in Quebec Sign Language.
Sign Language Studies 24: 99-139.
Boyes Braem, Penny and Rachel Sutton-Spence (eds.).
2001 The hands are the head of the mouth - the mouth as articulator in
sign languages. Hamburg: Signum Press.
Branchini, Chiara and Caterina Donati
in press Italian Sign Language relatives: correlatives, internally headed,or
free relatives? In Correlatives. Theory and typology, Aniko Lipták
(ed.). Amsterdam: Elsevier.

Sources of variation in sign language structure 29
Brentari, Diane
1998 A prosodic model of sign language phonology. Cambridge, MA:
MIT Press.
Byun, Kang Suk
2004 Gender marking in South Korean Sign Language. Paper presented at
the Max Planck Institute for Psycholinguistics, Nijmegen, July 2004.
Cecchetto, Carlo, Carlo Geraci and Sandro Zucchi
2006 Strategies of relativization in Italian Sign Language. Natural
Language & Linguistic Theory 24: 945-975.
Chen Pichler, Deborah
2001 Word order variation and acquisition in American Sign Language.
Ph.D. thesis, University of Connecticut.
Coerts, Jane
1992 Nonmanual grammatical markers: An analysis of interrogatives,
negations and topicalisations in Sign Language of the Netherlands.
Ph.D. dissertation, University of Amsterdam.
Crashborn, Onno
2001 Phonetic implementation of phonological categories in Sign
Language of the Netherlands. Utrecht: LOT Publication.
Emmorey, Karen
1996 The confluence of space and language in signed languages. In
Language and space, Paul Bloom, Mary A. Peterson, Lynn Nadel,
and Merrill F. Garrett (eds.), 171-209. Cambridge, MA: MIT Press.
Emmorey, Karen (ed.)
2003 Perspectives on classifier constructions in sign languages. Mahwah,
NJ: Lawrence Erlbaum.
Emmorey, Karen and Brenda Falgier
1999 Talking about space with space: Describing environments in ASL. In
Storytelling and conversation: Discourse in Deaf communities,
Elizabeth Winston (ed.), 3-26. Washington, DC: Gallaudet
University Press.
Engberg-Pedersen, Elizabeth
1993 Space in Danish Sign Language: The semantics and morphosyntax
of the use of space in a visual language. Hamburg: Signum Press.
Fischer, Susan D. and Yutaka Osugi
1998 Feature movement in WH-questions: Evidence from sign languages.
Poster presented at the 6th International Conference on Theoretical
Issues in Sign Language Research, Washington, DC.
2000 Thumbs up vs. giving the finger: Indexical classifiers in NS and
ASL. Paper presented at the 7th International Conference on
Theoretical Issues in Sign Language Research (TISLR), Amsterdam,
July 2000.

Pamela Perniss, Roland Pfau, and Markus Steinbach
30
Friedman, Lynn A.
1976 The manifestation of subject, object, and topic in the American Sign
Language. In Subject and topic, Charles N. Li (ed.), 125-148. New
York: Academic Press.
Hendriks, Bernadet
2004 An introduction to the grammar of Jordanian Sign Language. Salt
(Jordan): Al-Balqa Applied University.
Keenan, Edward L.
1985 Relative clauses. In Language typology and syntactic description.
Vol. II: Complex constructions, Timothy Shopen (ed.), 141-170.
Cambridge: Cambridge University Press.
Keller, Jörg and Trixi Rech
1993 Gegen Vorurteile: Gebärdensprache. Sprache und Kognition 12:
130-144.
Klima, Edward and Ursula Bellugi
1979 The signs of language. Cambridge MA: Harvard University Press.
Kooij, Els van der
2001 Weak drop in Sign Language of the Netherlands. In Signed
languages: Discoveries from international research, Valerie Dively,
Melanie Metzger, Sarah Taub, and Anne Marie Baer (eds.), 27-42.
Washington, DC: Gallaudet University Press.
Lehmann, Christian
1986 On the typology of relative clauses. Linguistics 24: 663-680.
Liddell, Scott K.
1978 Nonmanual signals and relative clauses in American Sign Language.
In Understanding language through sign language research, Patricia
Siple (ed), 59-90. New York: Academic Press.
2003 Grammar, gesture, and meaning in American Sign Language.
Cambridge: Cambridge University Press.
Liddell, Scott K. and Melanie Metzger
1998 Gesture in sign language discourse. Journal of Pragmatics 30: 657-
697.
Marsaja, I Gede and Ketut Kanta
2005 Kata Kolok: An introduction to a deaf village in Bali and its sign
language. Talk given at the Max Planck Institute for Psycho-
linguistics, Nijmegen, April 2005.
McBurney, Susan
2002 Pronominal reference in signed and spoken language: Are
grammatical categories modality-dependent? In Richard P. Meier,
Kearsy Cormier, and David Quinto-Pozos (eds.), 329-369.

Sources of variation in sign language structure 31
Meier, Richard P.
2002 Why different, why the same? Explaining effects and non-effects of
modality upon linguistic structure in sign and speech. In Richard P.
Meier, Kearsy Cormier, and David Quinto-Pozos (eds.), 1-25.
Meier, Richard P., Kearsy Cormier and David Quinto-Pozos (eds.)
2002 Modality and structure in signed and spoken languages. Cambridge:
Cambridge University Press.
Moravcsik, Edith A.
1978 Reduplicative constructions. In Universals of human language.
Vol.3: Word structure, Joseph H. Greenberg (ed.), 297-334.
Stanford, CA: Stanford University Press.
Morgan, Gary
1999 Event packaging in BSL discourse. In Storytelling and conversation:
Discourse in Deaf communities, Elizabeth Winston (ed.), 27-58.
Washington, D.C.: Gallaudet University Press.
Neidle, Carol, Judy Kegl, Dawn MacLaughlin, Benjamin Bahan and Robert G. Lee
2000 The syntax of American Sign Language. Functional categories and
hierarchical structure. Cambridge, MA: MIT Press.
Newport, Elissa and Ted Supalla
2000 Sign language research at the millenium. In The signs of language
revisited. An anthology to honor Ursula Bellugi and Edward Klima,
Karen Emmorey and Harlan Lane (eds.), 103-114. Mahwah, NJ:
Lawrence Erlbaum.
Nyst, Victoria
2007 A descriptive analysis of Adamorobe Sign Language (Ghana). Ph.D.
dissertation, University of Amsterdam.
Nyst, Victoria and Pamela Perniss
2004 Classifiers or generic directionals: Motion in Adamorobe SL and
German SL. Paper presented at the European Science Foundation
workshop “Modality effects on the theory of grammar: A cross-
linguistic view from sign languages of Europe”, Barcelona,
November 2004.
Padden, Carol A.
1988 Interaction of morphology and syntax in American Sign Language.
New York: Garland.
2002 How the alphabet came to be used in a sign language. In Progress in
sign language research: In honor of Siegmund Prillwitz, Rolf
Schulmeister and Heimo Reinitzer (eds.), 31-46. Hamburg: Signum.
Padden, Carol A. and David M. Perlmutter
1987 American Sign Language and the architecture of phonological
theory. Natural Language and Linguistic Theory 5: 335-375.

Pamela Perniss, Roland Pfau, and Markus Steinbach
32
Perlmutter, David M.
1992 Sonority and syllable structure in American Sign Language.
Linguistic Inquiry 23: 407-442.
Perniss, Pamela and Aslı Özyürek
in press Representations of action, motion, and location in sign space: A
comparison of German (DGS) and Turkish (TİD) Sign Language
narratives. In Signs of the time. Selected papers from TISLR 2004,
Josep Quer (ed.). Hamburg: Signum.
Petronio, Karen and Diane Lillo-Martin
1997 WH-movement and the position of Spec-CP: Evidence from
American Sign Language. Language 73: 18-57.
Pfau, Roland
1997 Zur phonologischen Komponente der Deutschen Gebärdensprache:
Segmente und Silben. Frankfurter Linguistische Forschungen 20: 1-
29.
2002 Applying morphosyntactic and phonological readjustment rules in
natural language negation. In Richard P. Meier, Kearsy Cormier, and
David Quinto-Pozos (eds.), 263-295.
Pfau, Roland and Josep Quer
2002 V-to-Neg raising and negative concord in three sign languages.
Rivista di Grammatica Generativa 27: 73-86.
Pfau, Roland and Markus Steinbach
2005a Backward and sideward reduplication in German Sign Language. In
Studies on Reduplication, Bernhard Hurch (ed.), 569-594. Berlin:
Mouton de Gruyter.
2005b Relative clauses in German Sign Language: Extraposition and
reconstruction. In Proceedings of the North East Linguistic Society
(NELS 35), Vol. 2, Leah Bateman and Cherlon Ussery (eds.), 507-
521. Amherst, MA: GLSA.
2006 Pluralization in sign and in speech: A cross-modal typological study.
Linguistic Typology 10, 135-182.
Pizzuto, Elena and Virginia Volterra
2000 Iconicity and transparency in sign languages: A cross-linguistic
cross-cultural view. In The signs of language revisited: An anthology
to honor Ursula Bellugi and Edward Klima, Karen Emmorey and
Harlan Lane (eds.), 261-86. Mahwah, NJ: Lawrence Erlbaum.
Sandler, Wendy
1989 Phonological representation of the sign. Linearity and nonlinearity
in American Sign Language. Dordrecht: Foris.
1996 Representing handshapes. In International review of sign linguistics,
William H. Edmondson and Ronnie B. Wilbur (eds.), 115-158.
Mahwah, NJ: Lawrence Erlbaum.

Sources of variation in sign language structure 33
Schembri, Adam
2003 Rethinking ‘classifiers’ in signed languages. In Karen Emmorey
(ed.), 3-34.
Schick, Brenda
1990 Classifier predicates in American Sign Language. International
Journal of Sign Linguistics 1(1): 15-40.
Schreurs, Linda
2006 The distinction between formally and semantically related noun-verb
pairs in Sign Language of the Netherlands (NGT). Master’s thesis,
University of Amsterdam.
Smith, Wayne H.
1990 Evidence for auxiliaries in Taiwan Sign Language. In Theoretical
issues in sign language research. Vol.1: Linguistics, Susan D.
Fischer and Patricia Siple (eds.), 211-228. Chicago: University of
Chicago Press.
Stokoe, William C.
1960 Sign language structure. An outline of the visual communication
systems of the American Deaf. Studies in Linguistics Occasional
Papers 8. Buffalo: University of Buffalo Press.
Supalla, Ted
1986 The classifier system in American Sign Language. In Noun classes
and categorization, Colette Craig (ed.), 181-214. Amsterdam:
Benjamins.
Supalla, Ted and Elissa L. Newport
1978 How many seats in a chair? The derivation of nouns and verbs in
American Sign Language. In Understanding language through sign
language research, Patricia Siple (ed.), 91-132. New York:
Academic Press.
Sutton-Spence, Rachel and Bencie Woll
1999 The linguistics of British Sign Language. Cambridge: Cambridge
University Press.
Talmy, Leonard
2003 The representation of spatial structure in spoken and signed
language. In Karen Emmorey (ed.), 169-196.
Tang, Gladys
2006 Questions and negation in Hong Kong Sign Language. In: Ulrike
Zeshan (ed.), 198-224.
Uyechi, Linda
1996 The geometry of visual phonology. Stanford: CSLI Publications.
Woll, Bencie
2003 Modality, universality and the similarities among sign languages: an
historical perspective. In Cross-linguistic perspectives in sign

Pamela Perniss, Roland Pfau, and Markus Steinbach
34
language research. Selected papers from TISLR 2000, Anne Baker,
Beppie van den Bogaerde, and Onno Crasborn (eds.), 17-27.
Hamburg: Signum.
Zeshan, Ulrike
2000 Sign language in Indo-Pakistan: A description of a signed language.
Amsterdam: Benjamins.
2003 ‘Classificatory’ constructions in Indo-Pakistani Sign Language:
grammaticalization and lexicalization processes. In Karen Emmorey
(ed.), 113-141.
2004a Hand, head, and face: Negative constructions in sign languages.
Linguistic Typology 8: 1-58.
2004b Interrogative constructions in sign languages: Cross-linguistic
perspectives. Language 80: 7-39.
Zeshan, Ulrike (ed.)
2006 Interrogative and negative constructions in sign languages.
Nijmegen: Ishara Press.
Zwitserlood, Inge
2003 Classifiying hand configurations in Nederlandse Gebarentaal.
Utrecht: LOT.