Johnston, van Roekel & Schembri (2015) On the conventionalization of mouth actions in Auslan...
Transcript of Johnston, van Roekel & Schembri (2015) On the conventionalization of mouth actions in Auslan...
1
TITLE PAGE
ON THE CONVENTIONALIZATION OF MOUTH ACTIONS IN AUSLAN
(AUSTRALIAN SIGN LANGUAGE)
Trevor Johnston, Macquarie University, Sydney, Australia
Jane van Roekel, Macquarie University, Sydney, Australia
Adam Schembri, La Trobe University, Melbourne, Australia
SHORT TITLE/RUNNING HEAD
Mouth actions in Auslan
ACKNOWLEDGEMENTS
Research for this study was supported by the Australian Research Council, grant #DP1094572 to Professor
Trevor Johnston. In addition to the annotations created by the three authors of this paper, we also wish to
acknowledge the work of students and research assistants who have also contributed to annotating the
Auslan Corpus. In chronological order since 2005 (most recent first): Lori Whynot, Donovan Cresdee,
Christopher Hansford, Ben Hatchard, Michael Gray, Gabrielle Hodge, Lindsay Ferrara, Julia Allen, Gerry
Shearim, Karin Banna, Louise de Beuzeville, Dani Fried, and Della Goswell. This study grew out of a pilot
BA honours research project by Jane van Roekel, supervised by Trevor Johnston.
LEAD AUTHOR CONTACT DETAILS
Professor Trevor Johnston
Department of Linguistics, Building C5A, Room 546
Macquarie University, North Ryde, New South Wales, Australia, 2109
2
ABSTRACT
This study investigates the conventionalization of mouth actions in Auslan (Australian Sign Language).
Signed languages were once thought of as simply manual languages because the hands produce the signs
which individually and in groups are the symbolic units most easily equated with the words, phrases and
clauses of spoken languages. However, it has long been acknowledged that non-manual activity, such as
movements of the body, head and the face play a very important role. In this context, mouth actions that
occur while communicating in signed languages have posed a number of questions for linguists: Are the
silent mouthings of spoken language words simply borrowings from the respective majority community
spoken language(s)? Are those mouth actions that are not silent mouthings of spoken words conventionalized
linguistic units proper to each signed language, culturally linked semi-conventional gestural units shared by
signers with members of the majority speaking community, or even gestures and expressions common to all
humans? We use a corpus-based approach to gather evidence of the extent of the use of mouth actions in
naturalistic Auslan—making comparisons with other signed languages where data is available—and the
form/meaning pairings that these mouth actions instantiate.
Keywords: mouthings, mouth gestures, non-manuals, constructed action, Auslan, Australian Sign Language
3
ON THE CONVENTIONALIZATION OF MOUTH ACTIONS IN AUSLAN
(AUSTRALIAN SIGN LANGUAGE)
INTRODUCTION
In this study, we report on the distribution of different types of mouth actions in Auslan (Australian Sign
Language), making comparisons with other signed languages where possible. We describe the characteristics
of mouth actions in Auslan and quantify the alignment of the attested form/meaning pairings found in a
recently created linguistic corpus of the language, i.e., a dataset which is representative of native and near-
native signer usage across a number of text-types, linguistically annotated, and machine readable
(searchable). We do this in order to gauge the extent and degree of language-specific conventionalization of
these actions. Mouth actions that occur while communicating in signed languages have posed a number of
questions for linguists: Are the silent mouthings of spoken language words simply borrowings from the
respective majority community spoken language(s)? Are those mouth actions that are not silent mouthings of
spoken words conventionalized linguistic units proper to each signed language, culturally linked semi-
conventional gestural units shared by signers with members of the majority speaking community, or even
gestures and expressions common to all humans? Our thesis is that, though there are pairings that suggest
some conventionalization of usage, much of the contribution mouth actions make to individual signs or
signed utterances may not be, unlike what has been claimed in the literature, specific to any particular signed
language or even signed languages generally. Rather, many mouth actions appear to be either part of the oral
and facial expression repertoires used by all humans—moderated by local cultural norms and, for hearing
people, constrained by the use of speech—or are co-articulated mouth patterns derived from the relevant
ambient spoken language.
The relevance of this study for our understanding of language is three-fold. First, it demonstrates the
importance of complementing intuitions and elicited datasets in signed language research with corpus-based
usage data: these cast doubt on earlier claims of high degrees of linguistic conventionalization, even
obligatoriness, in this area. Second, it underlines the complex relationship between signed and spoken
languages in many signing communities: the important role of mouthing must be acknowledged. Third, there
is a deep synergy between language in spoken and signed modalities: both use oral and manual gestures.
BACKGROUND
4
Signed languages were once thought of as essentially manual languages because the hands produced the
basic symbolic units of these languages (i.e., the signs) and these were readily equated with the words of
languages—and the phrases and clauses that words are the constituents of—with which linguists were
already familiar. Very quickly in the linguistic study of signed languages, however, it was realised that non-
manual activity which involves movements of the body, head and the face (especially involving the upper
face or eyebrows) play an important role in understanding the grammar of these languages (Baker & Padden
1978; Liddell 1978).
It has also long been known that there are many distinct signed languages. Most deaf signed language
using communities around the world use historically unrelated languages that are, for the most part, mutually
unintelligible. Despite these differences, signed languages across the world nonetheless appear to have many
features in common (Newport & Supalla 2000; Woll, Sutton-Spence, & Elton 2001), partly because their
visual-spatial modality allows room for meaningful uses of space, simultaneity and, in particular iconicity
(e.g., Cuxac 1999, 2000; Taub 2001; Perniss, Thompson & Vigliocco 2010; Meir, Padden, Aronoff &
Sandler 2013), and partly because the sociolinguistic situation of deaf signing communities around the world
is similar (e.g., Van Herreweghe & Vermeerbergen 2004; Lucas 2006; Quinto-Pozos 2007). Importantly, the
spatial, temporal and visual representational affordances of gesture-based face-to-face languages appear to
have been spontaneously and independently exploited in similar ways time and time again in different signed
languages. For example, the use of tilts, shakes and nods of the head, the raising and lowering of the eye-
brows, and the widening or narrowing of the aperture of the eyes have been observed repeatedly and their
functions in the grammars of signed languages have been described in varying detail with different amounts
of empirical support. That is, despite strong claims that some of these are formal syntactic markers of
relative clauses, content questions, conditionals and so forth, debate has continued not only on their functions
but also on whether they are syntactic or prosodic, and their degree of conventionalization or
grammaticalization in particular languages (cf. Pfau & Quer 2010).1
The mouth is prominent site of non-manual activity and movements of the mouth are an obvious
accompaniment to manual signing. The linguistic status of mouth actions in signed languages, like other
non-manuals, is also a question of debate (e.g., Vinson, Thompson, Skinner, Fox, & Vigliocco 2010).
1 See for example (in chronological order): Liddell 1980; Baker-Shenk 1983; Aarons, Bahan, Kegl, & Neidle 1992;
Johnston 1992; Nespor & Sandler 1999; Sandler 1999a, 1999b; Wilbur & Patschke 1999; Neidle, Kegl, Maclaughlin, Bahan, & Lee 2000; Tang 2006; Sandler & Lillo-Martin 2006; Zeshan 2006; Dachkovsky & Sandler 2009; Vos, Kooij, & Crasborn 2009; Wilbur 2009; Pfau & Quer 2010; Sandler 2010; Elliot & Jacobs 2013.
5
Previous research
Non-manual activity that is centred on the mouth (mouth actions) has been documented for a number of
signed languages and appears to be a characteristic of all those studied to date. The least problematic in
terms of categorization and identification is the class of mouth actions that are transparently complete or
partial silent articulations of the spoken words of the ambient spoken language. The remaining mouth actions
have the common characteristic in that they are not related to the articulation of spoken words, yet, at the
same time, they appear to fulfil a variety of functions which are not well understood. Since the publication of
a major overview of mouth actions in signed languages edited by Boyes Braem and Sutton-Spence (2001)
these two general types have usually been referred to as mouthings and mouth gestures respectively.2
An early issue of interest was the proportion of signs found to co-occur with mouth actions, especially
mouthings, in various signed languages (e.g., Schermer 1990). This was partly in response to well known
early claims that American Sign Language (ASL) used very limited amounts of mouthing when compared to
other signed languages (Padden 1980, Baker-Shenk 1983, Boyes Braem 2001). Another area of debate,
besides frequency, concerned the status of mouthings: were they an integral part of a signed language or
were they marginal, stemming from language contact or borrowing? Addressing both these questions
involved describing what mouthing and mouth gestures did and thus categorizing them into general types.
More recent publications by Fontana (2008), Lewin and Schembri (2011), Bank, Crasborn, and van Hout
(2011, 2013, submitted), and Sandler (2012) on mouth actions in Italian, British, Netherlands, and Israeli
signed languages respectively give a good overview of the research, particularly that since 2001.
Mouthings
Briefly, research has shown that mouthings frequently accompany manual signs in many signed languages,
including ASL,3 and there is some evidence—though the datasets have never been very large or varied—that
the rate varies according to text-type. They occur very frequently with fingerspellings. Mouthings have been
shown to occur more with nouns and morphologically simple verbs than with morphologically complex
2 For earlier research on this topic see Anderson and Reilly (1998) for ASL; Engberg-Pedersen (1993) for Danish SL
(henceforth we will use the abbreviation SL for “Sign Language” when naming a particular signed language), Ebbinghaus and Hessmann (1990) for German SL, Schermer (1985) for Netherlands SL, Schroeder (1985) for Norwegian SL, Bergman (1984) for Swedish SL, Vogt-Svendsen (1983) for Norwegian SL, Lawson (1983) for British SL, and Liddell (1978) for American SL. The Boyes Braem & Sutton-Spence volume included papers relating to British SL, Finnish SL, German SL, Indo-Pakistani SL, Italian SL, Netherlands SL, Norwegian SL, Swedish SL, and Swiss German SL. Some of
these papers are cited individually elsewhere in this study. 3 Nadolske & Rosenstock (2007:35) report that “Our finindgs indicate that—contrary to what has been claimed in the
literature—mouthings contribute significantly to the form and semantic aspects of ASL, in a similar manner to what has been claimed for European sign languages.”
6
signs such as indicating verbs (also known as agreement and spatial verbs) or with depicting signs (also
known as classifier signs), both of which may have aspects of their handshape, movement and/or location
modified meaningfully. Mouthing has been shown to add meaning to some signs by indicating a more
specific reading of a sign, e.g., the Auslan sign SPOUSE with the English mouthings ‘wife’ or ‘husband’
(Johnston & Schembri 2007). Indeed, for Schermer (2001) in these type of environments mouthing is often
obligatory in Netherlands SL (NGT). A mouthing can even add independent semantic information. Vogt-
Svendsen (2001) cites the example of the Swedish word for ‘red’ being mouthed along with the Swedish SL
sign PULL-OVER to mean ‘a red pull-over’. Although mouthings are often closely temporally aligned with
their co-articulated sign, they may be stretched, reduced, or repeated to maintain an alignment with the
duration and rhythm of the manual sign, especially if the sign has itself been modified (Fontana 2008).
Finally, the mouthing itself may spread regressively or progressively to adjacent signs (Crasborn, van der
Kooij, Waters, Woll, & Mesch 2008; Bank et al. 2013).
Mouth gestures
Mouth gestures are all other communicative mouth actions that are not mouthings. They have been described
as mouth actions that ‘do not derive from spoken language’ (Boyes Braem & Sutton-Spence 2001).
However, this needs qualification. It has yet to be shown that signers and speakers in the same speech
community and/or culture do not share any of the same or similar mouth action repertoires. Though some
research has been conducted on facial expressions that occur when speaking, as correlates of intonation and
stress or ‘audio-visual prosody’ (Krahmer & Swerts 2009), little or no research has been conducted on mouth
gestures made in face-to-face communication especially at moments when speakers are not actually
speaking but perhaps gesturing. Thus the extent to which the mouth gestures are actually sourced from or
related to the predominantly spoken language community in which signers are invariably embedded is as yet
unknown.
Some mouth gestures look like voiceless syllables. These syllabic mouth gestures are usually described
using letters, e.g., PAH, AP, THAM, WOOF, POW. The manual signs they co-occur with have been called multi-
channel signs because the mouth gestures add specific semantic components to the sign, e.g., Lawson (1983)
claimed that PAH means ‘success’ in British Sign Language (BSL) and is obligatorily produced with a sign
meaning ‘success’, ‘at last’, or ‘finally’. Woll (2001) proposed an alternative theory in which these mouth
7
movements echo hand movements, e.g., when a manual sign includes an abrupt closing of the hand, the
mouth will parallel this by snapping shut. Woll terms this ‘echo phonology’ as the mouth ‘echoes’ the
movement of the hand. Under this theory, these mouth gestures are devoid of semantic content but match
classes of manual movement such as opening, closing and twisting, e.g., PAH accompanies signs where the
handshape opens or the hands separate (often twisting apart).
Other mouth gestures are thought to act as morphemes with an adverbial or adjectival function (Liddell
1980). The classic example, given in Liddell for ASL and supported by other ASL researchers (McIntire &
Reilly 1988), and identified for other signed languages (Crasborn et al. 2008) including BSL (Lewin &
Schembri 2011) and Auslan (Johnston & Schembri 2007), is the protruding tongue (often written TH) to
mean something like ‘carelessly’. Other modifying ‘adverbial’ mouth gestures identified for some signed
languages have included EE for ‘effort’ (showing the teeth in a smile-like action), and MM (similar to a lip
pout) for ‘effortlessness’. ‘Adjectivals’ have included PUFFED-CHEEKS for ‘large/long things’, and SUCKED-
IN-CHEEKS for ‘small things’ (illustrated in Table 3 below). Woll (2001) categorized the syllabic type of
mouth gestures echoes and this second type as adverbials. She called a third sub-type enactions (henceforth,
we will use our preferred term, enactments). In enactments, the mouth gesture is part of the representation by
the signer of an event—a type of role-playing or mimesis—in which the mouth was involved one way or
another.
In an early study of Auslan (Johnston 1992), it was observed that some mouth gestures appear to have a
prosodic role. The upper face (eyebrows, eyes) and the lower face (mouth, tongue, cheeks) appeared to make
different contributions to visual prosody in Auslan—the former analogous to pitch, the second analogous to
stress. In other studies, mouthings that spread over two or more signs (‘stretched’ mouthings) have been
noted for their stress-like qualities in Swiss German SL (Boyes Braem 2001) and for their effect of binding
constituents together into a prosodic word when spreading from host to clitic (Sander 1999c) or into even
larger prosodic units such as noun phrases or verb phrases (Boyes Braem 2001). The prosodic role of some
mouth gestures, not just stretched mouthings, will once again be raised here, later in this paper.
Since 2001, researchers have continued descriptive work on mouth actions as well as investigating mouth
gestures as a part of non-manual grammatical signals.4 For a study of spreading behaviour in mouthings in
4 The research literature published in English in reverse chronological order would include, but is not limited to: Sandler
(2012) for Israeli SL, Lewin and Schembri (2011) for BSL; Pfau and Quer (2010) for SLs generally; Sandler (2009) and Dachkovsky and Sandler (2009) for Israeli SL; Fontana (2008) for Italian SL; Nadolske and Rosenstock (2007) for
8
three signed languages, Crasborn et al. (2008) called mouthings M-type mouth actions; adverbials mouth
gestures were called A-type (and now also included adjectival mouth gestures as some modified nouns), and
echoes became part of a slightly broader E-type category (for ‘semantically empty’ following Woll (2001)).
Enactments were discriminated into two sub-types—those involving only the mouth in which the mouth
represents itself doing the action described by the sign, such as BITE, LAUGH or LICK (these were called 4-
type for ‘mouth for mouth’): and W-type (for ‘whole of face’) in which the mouth is simply part of a large
whole-of-face expression, e.g., an open mouth with wide eyes for surprise. Crasborn et al. note that W-type
mouth gestures are not part of any potential mouth-based semiotic system because any mouth action form is
linked to the whole face and consequently its interpretation is a function of the enactment and not a function
of a possible conventional value attributable to the mouth form.
Gestural versus linguistic
More recently, some researchers have focussed on the question of the gestural nature of some mouth actions
in that they have the same kind of relationship to the conventional manual signs of a signed language as do
manual gestures to the conventional spoken words of a spoken language (Pizzuto 2003). Fontana (2008),
drawing on the work of Kendon (2004, 2008) and McNeill (2000), makes the radical suggestion that all
mouth actions can be analysed this way. A similar but more conservative observation is made in Dachkovsky
& Sandler (2009) and Sandler (2009) who identify a category of gestural iconic mouth actions to distinguish
them from syllabic E-type (or ‘lexical’) mouth components and from the conventional adverbial and
adjectival A-type modifiers already identified by other researchers (which they do not question).
The iconic category comes from the work of McNeill (1992) who, drawing on the work of earlier
twentieth century researchers (e.g. Efron, 1972; Kendon, 1980), explored how gesticulation played an
important role in both clarifying meaning and structuring discourse in spoken language. He divided gestures
into four categories, deictic, iconic, metaphoric and beat gestures. Iconic and metaphoric gestures portray
visually some aspect of the referent. Iconic gestures are usually made for concrete objects and events/actions
with a representation that physically resembles them (objects typically through an image and events/actions
typically through a mimetic enactment). Metaphoric gestures are usually employed to refer to ideas and
abstract expressions.
American SL (ASL); Johnston and Schembri (2007) for Australian SL; Torigoe and Wataru (2002) for Japanese (Okinawa) home sign.
9
Sandler found that some mouth gestures in Israel SL are mimetic replicas (oral mimes) and others act as
iconic gestures (symbolic icons), e.g., a round bowling ball was represented by puffed cheeks, a zigzag shape
by a sideways mouth movement, and the whoosh of an object swung through the air by an exhaled ‘faa’.
Different signers have varying versions of these iconic mouth gestures so there is no standardised or
conventionalised form. These iconic mouth gestures fit in well with McNeil’s (1992) classification of
gestures versus linguistic forms. It should be remembered that Sandler (2012) agrees with most signed
language researchers that, despite their gestural origins, many other mouth gestures and facial signals are
fully language-dedicated gestures that are an integral part of the lexis and syntax of signed languages, i.e.,
they are linguistic.
These three understandings of the linguistic (or gestural) status of mouth actions in signed languages are
illustrated in Figure 1.
Figure 1 Three possible categorizations of mouth actions
THIS STUDY
We document the distribution of types of mouth actions in Auslan—making comparisons with other signed
languages where possible—and describe their overall characteristics in order to test the language-specific
conventionalization of these actions.5
Methodology
5 Research towards this study was partially funded by an Australian Research Council grant #DP1094572.
10
Fifty video texts were selected from the Auslan corpus for analysis. All signs and all mouth actions were
examined and all mouthings and mouth gestures were identified, categorized and annotated using ELAN
multi-media annotation software.6 Annotations were exported into a database spreadsheet for further
processing. We conducted a statistical analysis of this data using Rbrul, with a multivariate logistic
regression model being used to identify predictors of mouthing and mouth gestures in the data (Johnson
2009). Like the program GoldVarb, developed by Rand & Sankoff (1991), Rbrul can quantitatively evaluate
the influence of multiple factors on variation.
Data
The data in this study have been drawn from the Auslan corpus of native or near-native signers (for further
details see Johnston & Schembri 2006). For this study, 50 video clips were selected from the corpus,
representing 38 individuals and 3 text types (monologue, dialogue, and elicited). The total duration of the
data was 5 hours and 58 minutes, representing 17,002 manual sign tokens. The signed texts ranged from 1:32
to 38:30 minutes in length. The 50 video clips consisted of 25 monologues (these were narratives, all of
which involved retellings of two Aesop’s fables); 10 dialogic texts (free conversation or responses to a series
of interview questions); and 15 sessions of 40 elicited picture descriptions.
Annotations
The 50 texts had been previously annotated according to guidelines detailed in the Auslan Corpus
Annotation Guidelines7 and in Johnston (2010). They were selected because they had previously received the
most detailed annotations of all the 459 corpus annotation files, and would thus require the least amount of
additional work, besides the study-specific annotations, to complete. That is, the majority had been almost
completely annotated for glosses and translations, as well as for clause boundaries and periods of enactment
or constructed action (a term for all forms of enactment or mimetic behaviour in signed languages, compare
with constructed dialogue from Tannen, 1986), and tagged for the relationships between clauses (e.g.,
embedded, dependent), the part of speech of clause constituents, and the argument roles of signs within the
clause (e.g., agent versus patient, etc.). Part of speech tagging, particularly relevant to this study, cannot be
carried out without first identifying clauses. Some files already had mouthings identified. We thus only had
6 http://tla.mpi.nl/tools/tla-tools/elan/download/.
7 http://www.auslan.org.au/corpus/annotations, last updated June 2014.
11
to complete the incomplete files before then adding the study-specific annotations: namely, annotations for
the state of the mouth as (i) a mouthing, (ii) a mouth gesture (which were further specified for sub-type as
explained below), or (iii) a period of no mouth action.
Annotations were added to two glossing and two grammatical class tiers. A tier for each hand is required
because it is possible to articulate two signs at the same time (one on each hand) so each may need to be
glossed separately. When this happens, the grammatical class of each sign/hand may also be different and
thus two tiers need to be available. Annotations for mouth gestures were made on the ‘mouth gesture form’
tier (called MouthGestF), and on the ‘mouth gesture meaning’ tier (called MouthGestM). The annotations for
mouthings were made on the ‘mouthing form’ tier (called Mouthing) (Figure 2).
Figure 2 The ELAN tiers used to annotate mouth actions [SSNc2a00:00:29.000]
The annotation schema for mouthings is summarized in Table 1.
Table 1
The annotation schema for mouthings
M-type (mouthing) Annotation Examples
Complete articulation COMPLETE-WORD RACE, RABBIT, VILLAGE, FAR
Initial segment I(NITIAL) V(ILLAGE), SA(ME), DIFF(ERENT), SH(EEP)
Medial segment (ME)DI(AL) (NO)TH(ING), (RE)MEM(BER) , (B)E(ST)
Final segment (FI)NAL (SUCCESS)FUL, (FIN)ISH, (IM)PROVE. (TO)DAY
Initial & final segment only IN(I)TIAL F(INI)SH, D(EA)F, S(UC)CESFUL
‘suppressed’ articulation* (SUPPRESSED) (LADY), (HAVE)
unreadable* Unreadable
anticipatory spreading
(regressive mouthing) MOUTHING-regr
ID gloss PT:PRO1SG EXPLAIN Mouthing EXPLAIN-regr EXPLAIN
“I explained…”
delayed spreading
(progressive mouthing) MOUTHING-prog
ID gloss FINISH PT:PRO1SG Mouthing FINISH FINISH-prog
“….I finished”
* A ‘suppressed’ mouthing annotation is used in a few instances where the annotators are convinced there is underlying movement
congruent with articulating a word associated with a sign, but the mouth does not actually open, e.g. the ‘y’ of ‘lady’ when signing LADY.
They are identified to distinguish them from mouth gestures, e.g. a EE-like mouth gesture. Where annotators were certain a word was
being mouthed—there are articulatory motions—but were simply unable to lip-read it, it is annotated as unreadable.
Mouth gestures were divided into types that were based on Crasborn et al. (2008) but further sub-
categorised, as illustrated in Figure 3. We did this because our first attempt to annotate mouth actions using
12
the existing categories appeared inadequate—we believed finer discrimination was necessary in order to
properly identify and account for each type.
Figure 3 Types of mouth actions annotated in this study
For example, A-types were divided into two-subtypes: the adverbials (and adjectivals), as described in
Crasborn et al. (2008), and prosodic mouth gestures. We did this because the latter did not appear to modify
the meaning of co-occurring signs like other ‘traditional’ A-types. They appeared to have an emphatic role,
not unlike stress in speech. (For possible manual correlates of stress in signed languages generally, see
Wilbur’s (1999) study of stress in ASL). Their most important characteristic appears to be a tensed posture
of the mouth which is held, even if relatively briefly, without changing dynamically, rather than any precise
form of the mouth posture itself, e.g., the WIDE or ‘EE’ mouth gesture used with a pointing sign, as in the two
signs on the left of the bottom row in Table 2. This sub-type was identified so that they could be compared
and contrasted to the ‘traditional’ category of adverbial mouth gestures: were they really ‘morphemes’ being
added to signs (as the A-type category is meant to capture) or actually part of the prosody?
A-type mouth gestures were annotated using a coding schema adapted from Sutton-Spence and Day
(2001) in their study of BSL. The schema attempts to capture objectively and consistently major features of
mouth gestures (mouth-open/mouth-closed, teeth-visible/teeth-invisible, puffed-cheeks/no-puffed-cheeks,
tongue/no-tongue, narrow/wide, etc.). The adaptation was adequate for our purposes and had the advantage
of allowing for an easy and direct comparison with the BSL data (see Appendix).
Some typical examples of the most general mouth gestures are illustrated in Table 2. We usually refer to
the mouth gestures in this paper by the short-hand glosses, but the codes indicate the range of related forms
covered by a single gloss.
13
Table 2
Mouth gesture (MG) form codes and glosses used for typical exemplars
BLOW
air moves inwards or outwards through the lips which may be pursed or rounded
CN8, CN17, ON16-18
BOTTOM-LIP-OUT bottom lip is pushed forward, out
or up
CN3, CN20, ON11, ON14
DOWN the corners of the mouth are pulled down, mouth
can be open or closed, lips can be pressed
together, tense or relaxed CN4, CN22, ON4, ON9, ON15
LIP-CURL
top lip is pulled up on one or both sides, as in a sneer
CN1, ON5, ON10
LIPS-OUT lips pushed forward, as in a pout or
“shh”
CN11-14, CN16, ON16
LIPS-PRESSED (‘MM’) lips are pressed together but the mouth corners are
relaxed
CN5, CN6, CN21, CN23,
OPEN
mouth is open ON1-3
PUFF puffed cheeks
CP1-8
SLIGHTLY-OPEN mouth is slightly open
ON6, ON12
SUCKED-IN
cheeks are sucked inwards CN24
TONGUE (‘TH’) tongue pokes out or is visibly forward all OT codes & CN19
TRILL (‘BRRR’) lips vibrate
CN7, CN9-10, CN13-15, CN18, CP5,
WIDE (‘EEE’)
the corners of the mouth are pulled wide, mouth can be open or closed, lips can be pressed together, tense or relaxed CN2, ON7, ON8, ON13, ON14
Five sub-types of W-type mouth gesturers were also identified. First, mouth actions that were part of
involuntary or spontaneous expressions (or indexes) of the state of the mind of the signer (e.g., amused,
confused, concerned) were categorized as spontaneous W-type. These were not coded for the present study
14
because we do not consider them to be candidates for inclusion in any semiotic system that is potentially
language-specific. Second, expressions or subtle meta-comments about what the signer is signing (i.e., the
signer’s attitude) were categorized as editorial W-type expressions. An example would be the use of LIP-
CURL (Table 2) together with furrowed brow, squinting of the eyes and a backward, recoiling, movement of
the head while producing HE SAY after signing YOU FAT (i.e., YOU FAT, HE SAY) to convey disapproval of the
overly direct and thus inappropriate comment being quoted. They were identified in order to gauge their
overall role in Auslan and, especially, how they differed from the A- or E-type mouth gestures. Third, mouth
gestures that were part of enactments were categorized as constructed action W-types. They involved the
whole face and the meaning was dependent on this context. If the constructed action involved the production
of one of the common recurring mouth gestures (as described above), the appropriate gloss for that was also
added to the annotation. The fourth sub-type were congruent W-types. These were expressions which carry
very little modifying communicative load and were, in fact, little more than a ‘congruent default expression’
i.e., they simply matched the semantics of the lexical sign, such as smiling while signing HAPPY, or looking
sombre when signing FUNERAL. These whole face affective expressions often appeared when particular
semantically appropriate lexical signs were cited. This phenomenon has led some researchers to claim they
are obligatory conventional non-manual components. However, though it would be decidedly odd to sign
HAPPY in isolation with a negative or sad expression (just as it would be odd to read out the English
headword depression from an English dictionary in a joyous manner), this does not mean the expression is
lexically specified merely because of a sympathetic ‘collocation’ of the congruent expression. Systematic
annotation of text quickly revealed that signs of this type can be signed with a neutral expression. We
identified these separately in order to quantify and compare them to any similar mouth gestures that do not
involve the whole face (if any) and to test if they really are obligatory. The fifth subtype, adverbial
expressive W-types, were clearly intended to modify and add meaning to the manual sign(s), unlike
congruent W-types, e.g., signing DANCE while performing it in a ‘lively’ or ‘energetic’ fashion with an
overall facial expression of enjoyment. However, they were not classified along with other A-types because
they were not restricted to the mouth and were strongly enacting, i.e., one way or another the signer was
engaging in constructed action. The annotation schema we used for mouth gestures is summarized in Table
3.
15
Table 3
The annotation schema for mouth gestures
Mouth gesture MouthGestF tier begins with MouthGestM tier contains
E-type (echo or empty) SYLL:GLOSS (= Syllable) various meanings as needed temporary tag tier: -IM (imagistic), -MI (mimetic), -ME (metaphorical)
A-type (modifying)
prosodic GLOSS/CODE(H) (H = held) (see Table 3)
meaning glosses: ACTIVITY, EMPHASIS
prosodic (non-specific) No annotation Tag tier: -MH (mouthing held)
adverbial GLOSS/CODE (see Table 3)
meaning glosses: LARGE-AMOUNT, CARELESS, UNPLEASANT, SMOOTH, EASE, EFFORT, SMALL-AMOUNT
temporary tag tier: -IM (imagistic), -MI (mimetic), -ME (metaphorical)
4-type (mouth for mouth) CMO (= Congruent Mouth Only) ENACTMENT
W-type (whole-of-face)
spontaneous no annotation
editorial COMMENT no further annotation or various meanings as needed
CA (Constructed Action) CA: (= Constructed Action) no further annotation or various descriptions as needed
CA using an A-type CA:GLOSS/CODE (Table 3) add after the CA: the A-type mouth gesture gloss/code
congruent CWF (=Congruent Whole Face) meaning glosses: EXPRESSION, ENACTMENT, EMPHASIS
adverbial expressive CA:ADV ( = Adverbial) EXPRESSION
Spreading mouth gesture ANNOTATION-cont on all subsequent co-articulated manual sign(s)
The time alignment of mouth action annotations When a mouth gesture is clearly articulated across 50% of
the duration of an adjacent sign, it is annotated separately on each co-articulated sign it spreads over with –
cont (for ‘continued’). When a mouthing begins before its target sign it is marked as –regr (for ‘regressive’)
on the sign during which it begins; where a mouthing persists beyond its target sign it is marked as –prog
(for ‘progressive’). Again, a 50% spreading criteria is applied. (The ‘target’ of a mouthing can usually be
easily identified but mouth gestures need not have an unambiguous single target, so they are simply marked
as continued from the sign in which they begin.) Only significant spreading is annotated in our dataset—the
exact onset and offset time of mouth gestures is not a focus of this study (other studies have explored this,
see Sandler 1999b; Crasborn et al. 2008; Bank et al. 2013).
Results
Overall distribution
About one in five signs were not accompanied by any noteworthy mouth activity, which is not unlike other
signed languages described to date (Table 4).
Table 4
Mouth action rates in Auslan compared with other signed languages*
Auslan BSL NGT SSL
16
Mouth actions 77% 71% 65% 90%
No mouth actions 23% 29% 35% 10%
* Data are taken from Crasborn et al. (2008).
The Auslan data also show that about half of all manual signs were accompanied by mouthing, and that
mouth gestures are relatively infrequent, especially 4-type, A-type and E-type (Figure 4).
Figure 4 Rate of mouth actions with manual signs
Approximately 73.6% of all mouth actions (i.e., excluding sign tokens with no mouth action) are, in fact,
mouthings, which is higher than in previous descriptions of other signed languages, although similar to the
recent corpus-based study of NGT (NGT-2 in Table 5) which also included several text-types.
Table 5
Each mouth action type as a % of all mouth actions compared to other signed languages*
Auslan BSL NGT SSL HKSL NGT-2
M-type 73.6% 51% 39% 57% 35% 80%
A-type** 8.3% 21% 30% 14% 17.5% n/a
E-type 0.5% 2% 8% 7% 9.5% n/a
4-type 0.5% 6% 6% 6% 2.5% n/a
W-type 17.1% 20% 17% 16% 36% n/a
* NGT-2 data from Bank et al. (2011, 2013), HKSL data from Siu Wai-yan (2007) ** Our W-type (adverbial expressives) are included for comparion because they appear most likely to have been coded as A-type by the other researchers.
Some of the differences in this data may be due to the finer discrimination of sub-types of the major
categories in the Auslan data, particularly the identification and annotation of mouthings. In our experience,
mouthings—whether complete or incomplete—are often not noticed on first viewing, or are often mistaken
for mouth gestures, by researchers for whom Auslan is a second language. They may only know the
17
‘standard mouthing’ (see below), or may not realize that some signers actually have larger English
vocabularies than they credit them with. They are thus not primed to see a mouthing.
Equally important are the single genre and small datasets that have been used. For example, the large
number of 4-types in the three European signed languages is almost certainly due to the fact that these
narratives involve a wolf, or a dog and a bone and present multiple opportunities for this kind of
representation. This bias was no longer present in the NGT-2 data.
Distribution by sign type
Studies of lexicology (Johnston & Schembri 1999; Brentari & Padden 2001; Cormier, Schembri & Tyrone,
2008), lexical frequency (Morford & MacFarlane 2003; McKee & Kennedy 2006; Johnston 2012; Fenlon,
Schembri, Rentelis, Vinson, & Cormier 2014), and lexicalization and grammaticalization (Janzen 2012;
Johnston & Ferrara 2012) in signed languages have revealed that they consist of different types of signs (e.g.,
core lexicon, productive lexicon, and gestural elements). If we are to understand the grammatical
organization of these languages, we need to ensure we properly distinguish between the types of signs in our
analyses, and not be misled by over-simplistic or ad-hoc glosses. Briefly, signed languages use not only fully
conventional lexical signs (more or less the equivalent to the conventional words of spoken languages)8 and
gestures, but also signs that appear to be blends of gestures and lexical signs (pointing signs, indicating
verbs, depicting signs). As the data in Figure 5 show, different sign types have very different rates of co-
articulated mouth actions, even before we consider what grammatical role tokens of each type may play in
any given clause or utterance.
8 And thus not to be confused with ‘lexical’ in the sense of ‘content’ (as opposed to ‘function’) words. Fully lexical signs
include examples of both ‘content’ and ‘function’ signs. The distinction ‘content’ versus ‘function’ is covered in the next section dealing with mouthings and grammatical class.
18
Figure 5 Mouth actions by sign type (ranked by decreasing % of mouthing)
Fingerspelled signs are clearly the type of sign most strongly associated with mouthing, but several
individuals from the corpus have very low rates, e.g., one mouths with only 28% of her fingerspellings.
Depicting signs and gestures were the least likely type of sign to have a co-occurring mouthing, though
they were the most likely to co-occur with a mouth gesture, by a considerable degree.
Multivariatiate logistic regression analysis, using Rbrul, showed that sign type was a significant predictor
of the use of mouthing (p = <0.001). The results (Table 6) are presented as log odds, with positive values
indicating an increased likelihood to favour a particular behaviour, while negative values indicate an
increased likelihood to disfavour this behaviour. Fingerspelled items (3.339), lexical signs (1.146) and
possessive signs (0.896) all strongly favoured mouthing, while pronominals (–0.128), locatives (–0.481),
buoys (–0.548), determiners (–0.692), depicting signs (–1.282) and gestures (–2.249) all disfavoured
mouthing.
Table 6
Sign type and mouthing
Application value: presence of mouthing
All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouthing
Sign type Fingerspelling 3.339 797 93
Lexical signs 1.146 11657 66
Pointing signs: Possessives
0.896 139 64.7
Pointing signs: pronouns
–0.128 1628 42.4
19
Pointing signs: locatives
–0.481 169 23.7
Buoys –0.548 127 37
Pointing signs: determiners
–0.692 139 18.7
Depicting signs –1.282 939 14.5
Gestures –2.249 1403 8.3
An additional Rbrul run (Table 7) showed the sign type was also a significant predictor of the use of mouth
gestures (p = <0.001). Gestures (2.313) and depicting signs (1.517) strongly favoured the use of mouth
gestures, as did buoys to a lesser extent (0.636). Pronouns (–0.053), lexical signs (–0.078), locatives (–0.097)
and possessives (–0.326) disfavoured mouth gestures to a certain extent, whereas determiners (–0.901) and
especially fingerspelling (–3.011) strongly disfavoured them.
Table 7
Sign type and mouth gestures
Application value: presence of mouth gestures All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouth gestures
Sign type Gestures 2.313 1403 62.8
Depicting signs 1.517 939 47
Buoys 0.636 127 20.5
Pointing signs: pronouns
–0.053 1628 13.6
Lexical signs –0.078 11657 15.6
Pointing signs: locatives
–0.097 169 21.3
Pointing signs: possessives
–0.326 139 11.5
Pointing signs: determiners
–0.901 139 11.5
Fingerspelling –3.011 797 1
The type of sign least likely to have any kind of mouth action (mouthing or mouth gesture) were
determiners, closely followed by locatives. Figure 6 shows there are considerable differences in mouth action
rates depending on the primary functional role of the pointing signs. On the whole, non-possessive pointing
signs (pronouns, locatives and determiners in Figure 5) had the least mouth activity, and none were mouthed
in the majority of cases. Possessive pointing signs had relatively high rates of mouthing compared to other
pointing signs. First and second person points were much more likely to be mouthed than locatives or
determiners.
20
Figure 6 Mouth actions with pointing signs (ranked by decreasing % of mouthing)
An Rbrul analysis (Table 8) found that the type of pointing sign was a significant predictor of the use of
mouthing (p = <0.001), although many of the types were represented only by small numbers of tokens. First
singular (1.943) and second singular (1.481) possessives strongly favoured mouthing. Other pointing signs
which favoured the use of mouthing included third person plural (0.820) and first person plural (0.715)
possessives, together with second person (0.456) and first person singular (0.075) pronominals. The
following forms disfavoured the use of mouthing: third person plural pronouns (–0.053), locative pointing
signs (–0.246), third person singular possessive (–0.314), determiners (–0.615), third person singular
pronouns (–0.652) and buoys (–2.093).
Table 8
Pointing signs and mouthing
Application value: presence of mouthing All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouthing
Pointing sign type 1sg possessives 1.943 67 82.1
2sg possessives 1.481 10 80
3pl possessives 0.820 15 53.3
1pl pronouns 0.715 24 62.5
2sg pronouns 0.456 160 53.8
1sg pronouns 0.075 844 47.7
3pl pronouns –0.053 125 42.4
Locatives –0.246 154 22.7
3sg possessives –0.314 44 36.4
Determiners –0.615 139 18.7
3sg pronouns –0.652 459 26.8
21
Buoys –2.093 27 11.2
Distribution by grammatical class
Mouth action rates in Auslan varied according to the grammatical class of the accompanying manual sign
(Figure 7).
Figure 7 Mouth action rates by grammatical class (ranked by decreasing % of mouthing)
Lexicalised nouns showed more mouthing than other classes of sign and even depicting signs functioning as
nominals were more likely to co-occur with mouthing than depicting verbs. Verbs, plain or indicating, were
more likely to co-occur with mouth gestures than mouthing. Apart from nouns, some of the highest mouthing
rates are with numbers and function words (prepositions, auxiliaries, conjunctions and wh-question words
and wh-relativizers).
These observations were confirmed by Rbrul analysis (Table 9), with grammatical class a significant
predictor of mouthing (p = <0.001). We found that numerals (2.283), nouns (1.804) and prepositions (1.582)
strongly favoured mouthing, as did relativizers (1.294), auxiliaries (1.066), adjectives (0.938), conjunctions
(0.919), question signs (0.381). Adverbs only slightly favoured mouthing (0.076). Plain and indicating verbs
(–0.298), negators (–0.360), depicting nouns (–0.375), pronouns (–0.984), locatives (–1.067), determiners (–
1.259), interjections (–2.116) and depicting verbs (–3.207) all disfavoured mouthing.
Table 9
22
Grammatical class and mouthing
Application value: presence of mouthing All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouthing
Grammatical class Numerals 2.283 114 89.5
Plain nouns 1.804 1089 81.2
Prepositions 1.582 105 80
Wh-relativizers 1.294 13 76.9
Auxiliaries 1.066 81 72.8
Adjectives 0.938 327 67.6
Conjunctions 0.919 93 68.8
Wh-question signs 0.381 72 50
Adverbs 0.076 492 47
Plain & indicating
verbs
–0.298 1656 38.2
Buoys –0.360 2 50
Negators –0.375 47 42.6
Depicting nouns –0.677 66 27.3
Pronouns –0.984 464 30.2
Locatives –1.067 22 22.7
Determiners –1.259 98 20.4
Interjections –2.116 521 11.9
Depicting verbs –3.207 413 2.9
Grammatical class was also a significant predictor of the use of mouth gestures (p = <0.001) (Table 10).
Depicting verbs most strongly favoured mouth gestures (2.538) followed by interjections (2054), buoys
(1.868), negators (1.366), plain and indicating verbs (1.108), depicting nouns (1.101), locatives (0.257),
adverbs (0.192), and pronouns (0.097). Question signs were neutral (0.000), while the following signs
disfavoured mouth gestures: adjectives (–0.251), relativizers (–0.296), determiners (–0.467), conjunctions (–
0.910), nouns (–1.602), prepositions (–1.785), numerals (–2.455), and auxiliaries (–2.813).
Table 10
Grammatical class and mouth gestures
Application value: presence of mouth gestures All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouth gestures
Grammatical class Depicting verbs 2.538 413 75.8
Interjections 2.054 521 57.6
Buoys 1.868 2 50
Negators 1.366 47 40.4
Plain & indicating verbs
1.108 1656 41
Depicting nouns 1.101 66 45.5
Locatives 0.257 22 27.3
Adverbs 0.192 492 22.2
Pronouns 0.097 464 18.1
Wh-question signs 0.000 72 20.8
Adjectives –0.251 327 15.6
Wh-relativizers –0.296 13 15.4
Determiners –0.467 98 14.3
Conjunctions –0.910 93 7.5
Plain nouns –1.602 1089 4.9
Prepositions –1.785 105 3.8
Numerals –2.455 114 1.8
23
Auxiliaries –2.813 81 1.2
This distribution of mouthing over signs of different grammatical classes observed in the data was, once
again, comparable to other signed languages for which data is available (Table 11).
Table 11
Distribution of mouthing by grammatical class of manual sign
Auslan BSL NGT SSL ASL* HKSL IrishSL*
Noun 35.8% 40.3% 39.4% 42.5% 27.1% 84% 32%
Verb 26.9% 24.4% 30.6% 20.9% 15.4% 9% 23%
Adverbial 8.4% 6.4% 11.1% 10.1% 20.5% n/a n/a
Adjective 8.8% 12.3% 5.6% 11.2% 26.7% 2.5% 27%
Other 20.1% 16.7% 13.4% 15.2% 10.3% 4.5% 18%
* ASL and IrishSL figures calculated from data reported in Nadolske and Rosenstock (2007) and Militzer (2013)
The only true outlier in Table 11 is HKSL. Three of the signed languages from communities in which
English is the dominant spoken and written language (Auslan, BSL and ASL) are somewhat similar in
distribution with Auslan and BSL, unsurprisingly, very similar indeed. The fourth English-related signed
language, Irish SL, had no category of adverbials included in the analysis and hence the data are difficult to
compare. Given the distribution of other categories and given that the adjectival category actually could
include many adverbials, it too may be fairly similar to the other Western and English-related signed
languages in the final analysis.
Standard mouthings and lexical frequency
Our data confirms the long made observation that some signs appear to be consistently mouthed with the
same word, called a ‘standard mouthing’ (Bank et al. 2011).9 Of 1,314 fully lexical sign types in the dataset,
442 had a token count of 5 or greater. Considering only these—in order to reduce the bias of signs with very
few tokens—we found that only 38 had only one associated mouthing: MORE, HELP, EASY, RAM, HOW,
FATHER, NOW, BALL, ENGLISH, AFTER, DOG, FIRST, MELBOURNE, SHOULD, THAN, TIME, TRAIN, WITH,
WORLD, ACCESS, FIVE, LOVE, WASTE, WEEK, BEGIN, CITY, CLASS, DEAF-CLUB, FEW, GOVERNMENT, HAPPY,
NEW, NEW-ZEALAND, PROBLEM, RETURN, VERY. It was much more common, therefore, for lexical sign types
to co-occur with more than one type of mouthing. Nonetheless, one of the co-occurring mouthings was
usually more frequent that any of the others and could thus also be regarded as the standard mouthing. Not
9 The existence of standard mouthings is actually one of the most salient criteria for selecting an English word to act as
the ID gloss for lexical signs in the Auslan Corpus.
24
surprisingly these mouthings are usually of words semantically related to the core meaning of the sign (and
its standard mouthing), e.g., beer, pub, bar with BEER. Seven of the fully lexical sign types with multiple
mouthings actually had ten or more unique attested mouthings (SAME, GO, SAY, GROUP, REAL, AREA, LOOK).
Most of these mouthings represented wordings semantically related to the core meaning of the sign (e.g.,
true, truth, real, really, actually, sure, honestly with REAL; or same, like, too, also, but, mean, with SAME),
but they also included co-articulations in which the mouthing added additional information for which there
was no separate articulated sign, e.g., have with GROUP, meaning ‘there was/existed a group’; or I’m with
SAY, meaning ‘I said “I’m…”’ or people with LOOK, meaning “people looked” (cf. Vogt-Svendsen 2001).
The data also reveal that only a handful of fully lexical signs were always mouthed, i.e., almost all, even
those with only one associated mouthing, had tokens that also had either no mouthing whatsoever or were
accompanied with a mouth gesture.
Several signs were produced with mouthing at rates much higher than other signs of the same grammatical
class, e.g., HAVE and FINISH. Unlike most other verbs, HAVE is almost always produced with a mouthing
(92%), almost always with ‘have’ and by ‘had’ when not; and FINISH is accompanied by a mouthing (as
‘finish’) in 85% of tokens. These mouthing rates are quite high and are comparable to the identical signs
HAVE and FINISH in BSL which Sutton-Spence and Day (2001) report as ‘near obligatory’.
Of signs with a token count of 5 or greater in the dataset, 26 fully lexical signs never co-occurred with
mouthing (SPRINT, COINCIDENCE, SUPPOSE, TEASE, SURPRISE, OBJECT, INJECTION, DISAPPEAR, AMERINDIAN,
EXCELLENT, LEARN, DO-THERE, ELIMINATE, AGE-YEARS(ONE), ATTEND, GO-TRACE, MAYBE, AUTOMATIC,
COLLECT, DELICIOUS, DEMOTE, MEET-SEE, PATIENCE, PUSH-ASIDE, RUN, UNTIL).
In order to conduct an Rbrul analysis relating the use of standard mouthings and lexical frequency, we
eliminated signs that occurred five times or less, reasoning that a sign occurring at such low rates in the data
could not be said to have a standard mouthing due to the lack of sufficient tokens. We identified a subset of
lexical signs that appeared to occur with the same English mouthing 50% of the time or more, and we
categorised this subset as those with standard mouthings.
We then ran an Rbrul analysis, with lexical frequency as a continuous dependent variable, and mouthing
status as a fixed effect, and participant as a random effect (Table 12). We found that lexical frequency was a
significant predictor of the use of standard mouthing (p =<0.001), so that standard mouthings were favoured
with higher frequency (98.593) compared to non-standard mouthing (–31.812) and no mouthing (–66.780).
25
Table 12
Lexical frequency and standard mouthings
Application value: lexical frequency as a continuous variable All factor groups significant at p <.05
Factor Group Factor Log odds Tokens mean
Mouthing status Standard mouthing 98.593 510 420.857
Non-standard mouthing
–31.812 7175 291.372
No mouthing –66.780 3129 249.043
Characteristics of different types of mouth actions
M-types Insofar as can be discerned from lip-reading alone, most mouthings appeared to be fully articulated
(over 95%) (Table 13). Incomplete mouthings were overwhelmingly initial segments only, cf. NGT (Bank et
al. 2013). A small number of mouthing tokens (approximately 20) were not accompanied by any manual sign
yet they were clearly not redundant. They provided essential disambiguating or logico-cohesive information
to the utterance. They were conjunctions, prepositions or adverbial including ‘but’, ‘or’, ‘for’, ‘just’,
‘maybe’; sentence modifiers like ‘I don’t know’, ‘I think’; or other interactives like ‘no’, ‘yes’, ‘not true’.
Table 13
Type of mouthing
Degree of articulation Tokens
Full articulation 9197
Initial segment 289
Medial segment 16
Final segment 24
Initial & final segment only 24
Suppressed articulation 6
Unreadable 81
Total 9639
E-types Empty, echo or syllabic mouth gestures were rare in the data, but in all cases they had the minimal
effect of making the meaning of the co-occurring manual sign more emphatic. A full list of types and tokens
is given in Table 14.
Table 14
E-type mouth gestures in the dataset (N = 63)
Gloss MouthGestF Tokens Gloss MouthGestF Tokens
COINCIDENCE PAH 11 WELL-KNOWN PAH 1
REAL AP 7 DSS(H):ANIMAL-RUNNING PAH 1
26
FINISH.FINALLY PAH 5 ELIMINATE THAM 1
AUTOMATIC WOOF 4 SLEEP PAH 1
HAVE-NOT POOH 3 DSS(B):OBJECT-PASSES PAH 1
BAN AP 3 SPEED-DUST BOOM 1
PECULIAR PAH-PAH 2 DSM(B):ANIMAL-OVER-LINE PAH 1
LEARN-LESSON PAH 2 DSM(B):ANIMAL-STOPS PAH 1
VANISH AP 2 DIMINISH THAM 1
ZOOM-OFF POW 1 AFTERGLOW AM 1
RELIEVED PAH 1 FOUND-OUT PAH 1
UNUSUAL PAH 1 LEARN OOM_INHALE 1
REAL ALAM 1 ELIMINATE ALAM 1
EMPTY THAP 1 DSS(H):ANIMAL-RUNNING PAH 1
WITNESS PAH 1 TOO-LATE PAH 1
WIPE-UP ALAM 1 SPASMODIC AP 1
COHHLEAR-IMPLANT PAH 1 DSS(4):ASSEMBLY-LINE PAH 1
The number of different syllabic mouth patterns in this dataset is actually quite small, at 12 forms, and in
every case some discernible meaning is associated with the mouth gesture in addition to the underlying
emphasis (Table 15). These forms also have consistent, though admittedly general, meanings across usage
contexts, i.e., they are not semantically empty as such as suggested in the literature.
Table 15
E-type mouth gestures by associated meanings in context
Mouth gesture form Specific contextual meanings # echo
phonology # echo
metaphorical # NOT echo
# tokens total
PAH suddenly, quickly 25 4 2 31
AP suddenly, abruptly 8 5 0 13
WOOF without-impediment, automatic 0 0 4 4
ALAM all-gone, disappear from view abruptly
1 2 0 3
POOH nothing, negative, remove, blow away
0 3 0 3
THAM all-gone, disappear from view 0 2 0 2
PAH-PAH peculiar, strange 0 0 2 2
AM sudden, complete 1 0 0 1
BOOM all-gone, complete, energetic 1 0 0 1
OOM_INHALE close-shave, sharp, risky, dangerous 0 1 0 1
POW sudden, energetic 1 0 0 1
THAP all-gone, disappear from view
abruptly 0 1 0 1
Totals 37 18 8 63
The underlying emphasis appears to stem from the parallel reinforcement of the form of the manual sign by
the mouth gesture: the syllabic opening and closing of the mouth aligns itself with opening and closing of
handshapes, the striking of surfaces, or the sudden change in hand orientation (twisting, turning and
27
bending). These mouth gestures thus generally conform to the characterization of echo phonology proposed
by Woll (2001).
One syllabic mouth gesture—PAH—did not display in one instance the type of echoing predicted or
attributed to it. In this instance, PAH was aligned with the rapidly closing handshape in SLEEP, which is the
opposite of the opening movement it normally aligns with and ‘echoes’. Nonetheless, in all instances PAH
has to be said to be optionally ‘added’ because, again in contrast to some of the earlier literature (Sutton-
Spence & Woll 1999),10
signs in this class of ‘multi-channel signs’, such COINCIDENCE, FINISH, or REAL can
occur without this mouth gesture: of 19 instances of COINCIDENCE, 11 occur with PAH, 1 with A-type lips-
pressed, 2 with W-type expressions, and 4 with no mouth action at all; of 32 tokens of FINISH, 5 with PAH,
and 27 with mouthing; and of 179 tokens of REAL only 7 tokens occur with AP (137 tokens are mouthed).
Syllabic mouth gestures are attested across a variety of signed languages. According to Crasborn et al.
(2008) they accompany 1.3% of signs in BSL, 4.9% in NGT and 6.1% in SSL. In our dataset, they occurred
with only 0.4% of signs, a much lower rate than the BSL data.
Adverbial A-types Only a very small percentage of A-types are potential candidates for conventional mouth-
based adverbials (i.e., with their own dedicated or language-specific semantic meaning) that are applied as a
modifier to the co-occurring manual sign. The data suggest that combinations of manual signs and these
mouth gestures (with may occur with other non-manual elements) can produce specific meanings in
particular usage contexts, but that overall the semantic component of the mouth gesture is very broad. Only a
very small set of recurring semantic descriptors was needed to capture the apparent contribution of these
mouth gestures—LARGE-AMOUNT, CARELESS, UNPLEASANT, SMOOTH, EASE, EFFORT and SMALL-AMOUNT
(Table 16).
Table 16
Form/meaning pairings for adverbial A-type mouth gestures* (N = 187)
MouthGestF MouthGestM Specific contextual meanings Tokens
TONGUE CARELESS carelessly, easily, with ease, without regard, petulantly, with deliberate careless
enjoyment, reckless, slipshod, insouciant 109
TRILL EASE easily, unimpeded, with enjoyment 16
10
Woll (2001) does qualify this earlier claim by saying the mouth gesture is only obligatory in the citation form,
presumably recognizing that usage data may not support a stronger claim. However, it is not clear if this better accounts for the data. Many signs cited out of a usage context are accompanied by typical mouthings and mouth gestures, especially if they can have a range of meanings, as part of creating a more specific meaning for that form. Thus, it is not clear if the use of mouthing and mouth gestures to modify a sign’s meaning should be confounded with obligatoriness.
28
MouthGestF MouthGestM Specific contextual meanings Tokens
BLOW SMALL-AMOUNT little remaining, blown away, 12
BLOW SMOOTH smooth, unimpeded, quickly, ongoing 11
TRILL LARGE-AMOUNT large amount, a lot of, unimpeded, energetic, powerful, engine/machine-powered 8
TONGUE UNPLEASANT unpleasant, distasteful, bad 7
PUFF LARGE-AMOUNT large amount, a lot of, powerful 6
BOTTOM-LIP-OUT CARELESS careless, easily, without regard, petulantly, with deliberate careless enjoyment, reckless, slipshod, insouciant
3
LIPS-OUT EASE easily, without regard, petulantly, with enjoyment 2
LIPS-OUT SMALL-AMOUNT small amount, trivial, insignificant, nothing to worry about 2
LIPS-PRESSED EASE easily but deliberately, enjoyable 2
LIPS-OUT LARGE-AMOUNT large amount 1
PUFF CARELESS careless 1
TONGUE SMOOTH smooth 1
WIDE EFFORT effort 1
LIPS-PRESSED EFFORT effort 1
LIPS-OUT UNCLEAR n/a 1
SUCKED-IN SMALL-AMOUNT small amount 1
DOWN CARELESS careless 1
SLIGHTLY-OPEN EFFORT effort, concentration 1
* Again tokens with more than one possible descriptor were aggregated with a single descriptor that was the most
salient.
Prosodic A-types The majority of A-types appear to be prosodic in character, according to our definitions.
These mouth gestures tended to accentuate an action qua action, i.e., as an unfolding process, not unlike
Sandler’s ‘protracted action’ category of mouth gestures (2009). (She, however, analyses ‘protracted action’
expressions as conventional adverbial modifiers proper to Israeli SL, not mouth gestures.) We judged that
only two broad labels are needed to capture the effect of these prosodic mouth gestures: EMPHASIS and
ACTIVITY (Table 17). EMPHASIS was the broader default reading accounting for 71% of tokens, ACTIVITY
accounted for the force of the remaining 29%.
The actual mouth gesture used with each token varied considerably (and idiosyncratically) when collated
according to our form-based coding system. Nonetheless one underlying feature—a tensed mouth—did unite
most of them (all but TRILL, PUFF and BLOW). This supports earlier descriptions of Auslan in which a facial
‘grimace’ was said to indicate emphasis or stress (Johnston & Schembri 2007). If we use the broader general
form descriptors for A-type mouth gestures (which seem more appropriate and manageable given this
variability, see Table 2), we can see that the tensed mouth gestures account for 78% of the prosodic mouth
gestures. DOWN, WIDE, LIPS-PRESSED and TRILL account for 62% of EMPHASIS, and WIDE is the most
frequent for ACTIVITY.
29
Table 17
Form/meaning pairings for prosodic A-type mouth gestures*
MG form (= EMPHASIS) Tokens MG form (= ACTIVITY) Tokens
DOWN 88 WIDE 66
WIDE 78 TRILL 28
LIPS-PRESSED 65 SLIGHTLY-OPEN 22
TRILL 63 LIPS-PRESSED 22
SLIGHTLY-OPEN 38 DOWN 19
LIP-CURL 34 BOTTOM-LIP-OUT 13
PUFF 31 PUFF 8
BOTTOM-LIP-OUT 27 LIPS-OUT 8
OPEN 23 BLOW 4
BLOW 16 OPEN 3
LIPS-OUT 13 LIP-CURL 1
Totals 476 Totals 194
* A number of mouth gestures need to be translated from the formal
codes using more than one of these descriptors (e.g., “blow, lips-out”). In
these cases, the tokens were aggregated into the most salient descriptor with the same general meaning.
4-types The number of mouth for mouth gestures was extremely low. The token frequencies are
unremarkably linked to the narratives chosen for the re-tells or the elicitation materials. Insofar as these are
all mimetic of the mouth action involved in the process designated by the co-articulated manual sign, they
are also ‘iconic’.
Table 18
Mouth for mouth (N = 62)
ID-gloss Tokens ID-gloss Tokens ID-gloss Tokens
GRAZE 11 EAT 5 HEARING 2
CAPTURE 10 SHOUT 3 CATCH 1
YELL 10 LAUGH 2 DSS(5):ANIMAL-TEETH 1
AMERINDIAN 6 CHEW 2 BITE 1
SPEECH 5 ANGRY 2 CHATTER-BOX 1
W-types From the data presented in Figure 4 one can calculate that 71.4% of mouth gestures alone are of the
W-type. Of these we find that 42.6% are of the constructed action sub-type. There are very few of the other
types. Indeed, constructed actions represent 30.4% of all mouth gesture types (i.e., mouth actions excluding
mouthings). These were further coded on the MouthGestM tier as ENACTMENTS (global enactments of the
subject of the constructed action, including their facial expression), or EXPRESSIONS (presentations of just the
face expression of the subject during the constructed action). A large majority were of the second type. Many
ENACTMENTS and EXPRESSIONs used mouth actions found in the A-type mouth gesture repertoire to convey
30
similar meanings, e.g., TONGUE for ‘carelessly’. As a result, at minimum 10% of all putative adverbial A-
type mouth gestures in the dataset that clearly modify a co-articulated manual sign in some way are actually
part of enactments, i.e., constructed action W-types.
Iconicity
The gestural nature of some mouth actions has been recognized by all researchers. The very category of W-
types is intended to identify a large proportion of these. Indeed, most of the mimetic replicas and some of the
symbolic iconics that Sander (2009) identifies in Israel SL fall into this category. However, some mouth
actions that do not include the whole face, and thus fall into the A-type and E-type categories as coded in this
and other cited studies, also appear to be drawing on mimetic replicas or symbolic iconics—the form of these
mouth actions are related to their meanings. Consider the A-type puffed cheeks mouth action meaning
LARGE-AMOUNT, or the E-types in which the movement of the hands (which the mouth ‘echoes’) is itself
iconic, e.g., when the fingers close to represent something getting smaller and smaller in DISAPPEAR. So, in
order to assess the amount of gestural iconicity that may be present in our mouth action data, and evaluate it
in the light of Sandler’s (2009) observations, we identified those A- and E-type tokens in our data where
iconic motivation is present and identified its type (i.e., as imagistic, mimetic, or metaphoric). We could then
combine these token counts with W-types token count.
Among the 63 identified E-types, 18 were examples of echo phonology in which the form was an iconic
metaphor. Of the A-types, at least 82 (45%) of the 187 ‘traditional’ adverbial A-types (i.e., those not in the
sub-group prosodic) were iconic in some way (usually metaphoric). Consequently, approximately 40% (100
of 250 tokens) of A- and E-type mouth gestures appear to be symbolic iconics. The majority fell into two
main groups: those involving the puffing of the cheeks with or without the expulsion of air or the vibrating
of the lips and were associated with the intensification/augmentation; and those involving the pursing of the
lips or narrowing of the mouth, with or without the sucking in of the cheeks and the inhalation or expulsion
of air, and were associated with diminution/attenuation on the one hand, or care/precision on the other.
The forms the remaining prosodic A-types assume appear to be directly related to the stress or emphasis
they impart and are thus also non-arbitrary. Indeed, we suggest these prosodic mouth actions are expressions
deaf or hearing people spontaneously make when ‘doing or uttering with stress’ or imitating someone doing
so. They may be considered mimetic replicas of an action produced with stress or emphasis, if not actually
31
indexic reflexes (the actual manifestation of stressed production). If considered in this light, then in as many
as 80% of all A- and E-types, the form of the mouth action maps onto the meaning in an iconically motivated
way (imagistically, mimetically) which may also be metaphorical. Some of these may even be indexic. When
these are combined with W-types, fully 96% of all mouth gestures thus appear to be iconically motivated.
Spreading
Spreading behaviour in mouthings in the Auslan data appears less frequent than that reported for other
signed languages, despite the fact that the criteria we used for annotating spreading was the same as the other
projects (i.e., that the mouth action must spread over at least 50% of the duration of the adjacent sign). Once
again it is possible that different annotation practices may partially explain this (see Table 5). The spreading
behaviour of mouth gestures when considered alone is, however, comparable to some of the spreading rates
reported for the two of the other signed languages. By nature, mouth gestures are not as strongly aligned with
single signs, having a significant prosodic nature.
Table 19
Spreading data
Mouth action Tokens Percent with spreading
Auslan BSL NGT-1 SSL Auslan BSL NGT SSL NGT-2
M-type (mouthing) 9639 560 299 831 3.5% 25% 20% 12.5% 12.5%
Mouth gesture 3459 539 458 624 13.3% n/a n/a n/a n/a
E-type 63 20 58 99 0% 10% 5.3% 11.1% n/a
A-type 861 231 230 205 8.3% n/a n/a n/a n/a
4-type 62 63 45 87 0% n/a n/a n/a n/a
W-type 2472 225 125 233 15.7% n/a n/a n/a n/a
Nonetheless, the spreading behaviour itself is similar to that which has been described with other signed
languages, in terms of direction and source and target. There were examples of all of the types of spreading
mentioned in the literature: the most common with mouthings were progressive (88%) but some were
regressive (12%) (cf., 94.1% and 5.2% respectively for NGT-2); of these, the majority were from content or
lexical signs to function signs or to pointing or depicting signs, e.g. 49% spread over an adjacent pointing
sign (cf. 58.3% for NGT-2). The next most frequent were spreadings between lexical signs and the
fingerspellings that represent the English equivalent of the sign’s intended meaning (the target and source in
these examples is uncertain because fingerspellings appeared on either side of the lexical sign).
32
Variation
Individual variation There was enormous variation between individuals in the dataset in their use of mouth
actions. The rate of mouthing by individuals ranged from a low of 5% to a high of 82%. Mouth gesture rates
ranged from a low of 2% to a high of 43% (Figure 8). The mouthing rates for the six individuals in the BSL,
NGT and SSL datasets which are based on two signers from each language (Crasborn et al. 2008) have been
included in Figure 8 for comparison. (Because only mouthing rates were reported in Crasborn et al., the rates
for mouth gestures and no mouth actions are absent and there are white/empty bars above these data points in
the graph.)
Figure 8 Mouth action rates for individual signers
Overall there is a tendency for signers to use mouth gesture less frequently as mouthing rates increase. This
is not surprising: the more one mouths, the less opportunity one has to perform a mouth gesture.
Text-type variation Mouth action rates vary across text-types (Table 20). Although Sutton Spence and Day
found rates over 70% for some registers in BSL, their narrative samples had rates around 44%.
Table 20
Mouth action rates across text-types
No mouth action M-type E-type A-type 4-type W-type
Monologue (narratives, retell)
30.8 20.4 55.4 26.2 72.6 45.6
Dialogue (conversation, interview)
53.3 68.5 44.6 49.8 17.7 48.0
Elicited 15.8 11.1 0.0 23.9 9.7 6.4
33
(sentence elicitation, picture stimulus)
The fact that the individuals and the number of individuals represented in each text-type count is different
may account for some of the range, or lack of range, in variation. However, 11 individuals in the dataset
have examples of production in at least two text-types, and one has an example from each of the three text-
types (Figure 11). This data gives us confidence that the differences in Table 20 are not driven merely by
individual differences, which we still expect to be significant. The data appear to support the notion that rates
do vary with text-type. Nine of the eleven individuals show a difference in mouthing rates in the same
direction (namely more mouthings in dialogues, and more W-type in the monologues) between each of two
or more text-types.
Figure 9 Mouthing rates in different text-types with same participant
In a Rbrul analysis, text type was found to be a significant predictor of mouth action types (p = <0.001).
Monologues strongly favoured 4-type mouth (1.349) and E-type (0.976) mouth gestures, while weakly
disfavouring A-type (–0.111) and W-type (–0.299) mouth gestures, and particularly disfavouring mouthing
(–1.095) (Table 21). In contrast, dialogues favoured mouthing (0.801), a lack of any mouth actions (0.600),
W-type (0.460) mouth gestures, and disfavoured A-type (–0.455), E-type (–0.653) and 4-type (–0.752)
mouth gestures (Table 22).
Table 21
Monologues and mouth action types
Application value: use of mouth action types
34
All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouth action type
Monologues 4-type 1.349 62 72.6
E-type 0.976 63 55.6
A-type –0.111 884 46.5
W-type –0.299 2463 46.1
No mouth action –0.819 3939 30.8
Mouthing –1.095 9586 20.3
Table 22
Dialogues and mouth action types
Application value: use of mouth action types All factor groups significant at p <.05
Factor Group Factor Log odds Tokens % with mouth action type
Dialogues Mouthing 0.801 9586 68.6
No mouth action 0.600 3939 54.0
W-type 0.460 2463 47.6
A-type –0.455 884 27.9
E-type –0.653 63 44.4
4-type –0.752 62 17.7
Multivariate analysis of social factors We conducted multivariate logistical regression analyses using Rbrul
on the occurrence of mouthing and mouth gestures with the following social factors as independent
variables: signer’s sex, age, nativeness and region, with participant as a random effect in a mixed effect
model. Although participant variation was significant, we found none of the social factors were predictors of
the use of either mouthing or mouth gestures.
Discussion
Methodology
Type and size of datasets At 17,002 manual sign tokens, the Auslan dataset used in this study is much larger
and varied than previous studies of BSL, NGT, SSL and HKSL already cited, or even the next largest study
(at 11,905) recently completed which we have referred to as the NGT-2 dataset (Bank et al. submitted).
Previous researchers called for further research with significantly larger numbers of signers to help
distinguishing between variation between individual signers and differences between signed languages
(recall Figure 8). Our dataset enables us to see quite a wide distribution in the rate of mouth actions for
individuals, despite the fact that we have deliberately restricted our corpus to native or near-native signers.
One cannot fail to see from Figure 8 that if we were to randomly select two individuals from our dataset to
represent Auslan mouthing rates (or rather had only selected two individuals in the first instance), the results
35
would most likely have been quite misleading. Similarly, the spreading data for NGT-1 are quite different to
the spreading data for NGT-2.
It is relevant in this context to note that it has been taken as axiomatic by many signed language
researchers that almost all of the symbolic communicative behaviour of signing deaf people is ‘linguistic’ in
the narrow sense. However, this is actually simply a working assumption, in some cases driven by linguistic
theories which assume the autonomy of language from other aspects of cognition and communication. If
gesture plays a significant role in face-to-face communication (spoken or signed), then some symbolic
behaviour may not be linguistic in the sense of being part of a highly conventionalised, rule-governed system
in which most of the forms—in either primary modality—are actually language-specific. Possibilities in
wording and morpho-syntactic coding are often highly constrained by the very nature of linguistic systems,
i.e., some constructional schemas are obligatory in certain contexts and thus many aspects of linguistic
symbolic behaviour can be sampled from relatively small numbers of users precisely because of this.
However, if the substantive symbols are not actually ‘linguistic’ in the sense we have described, it is unlikely
that any single individual, or small sample of individuals, will provide data upon which can be generalized
core constructional schemas of the language. There is reason to believe that some aspects of signing
behaviour (like mouth actions) fall into this category.
It is for this very reason we have adopted a corpus-based approach in this study (and in other areas of the
description of Auslan). Nonetheless, one may reasonably contend that no datasets yet presented, including
our own, are sufficiently representative of native signer usage for our observations to be unlikely to be
revised in the future. This needs to be borne in mind.
Coding and categorization In a first parse of the data, A-type mouth gesture forms were annotated using the
glosses as shown in Table 2. When this first round of data was analysed, our results regarding form/meaning
pairings were no different than reported above and discussed below. Given the strong claims made in the
literature on the conventionalization of some mouth gestures, we considered that our coding and
categorization was perhaps too general to capture this, so we then used the Sutton-Spence and Day (2001)
schema, which we adapted (see Appendix), and returned to the primary data and re-coded the entire dataset.
The finer coding did not alter the outcome: there remained the range of variation in forms that we originally
observed, i.e., the more discrete and detail features did not align in any discernable pattern, and the relative
36
importance of the co-text and context of utterance in describing the meaning of these forms was also
unchanged.11
Categorizing mouth gestures is neither simple nor straightforward. Mouth gestures often overlap, e.g., if
the mouth is widened, with the lips closed, the lips will often be pressed together, or different lip
configurations can occur with a slightly open mouth. This is evident from an examination of the form-based
codes we used. Moreover, decisions about how to label the meaning of mouth gestures are even more
subjective and thus potentially open to re-interpretation by other researchers, even if reviewed by a native
signer (as they were). Inconsistencies could also arise if an annotator classifies a mouth action as a non-W-
type and fails to notice that the whole face is actually involved or that the mouth action is performed during a
period of constructed action. Again, one needs to keep this in mind.
Language contact and the conventionalization of mouth actions
Relationship of English mouthings to Auslan signs The majority of manual sign tokens (56.7%) co-occur
with mouthing (Figure 4). Given mouthing rates range from 6% to 84% (Figure 8), can it still be suggested
that mouthing is an obligatory formational component of some signs?
The strongest claim has been made for fingerspelled signs. Only some 6% of fingerspellings have no
mouthing (or mouth gesture) (Figure 5). Several individuals in the dataset, however, have very low rates of
mouthing with fingerspelling e.g., MVS who is from a multi-generational deaf family, mouths only 4 of 14
fingerspellings (a rate of only 28%). MVS is not difficult to read for a native signer and her signing style
should certainly not be considered atypical.12
The next strongest claim has been made for the mouthing co-occurring with so-called ‘ambiguous’ signs,
such as Auslan SPOUSE or NGT SIBLING.13
True, all 5 instances of SPOUSE in the dataset were indeed
mouthed ‘wife’ or ‘husband’ and it is clear that signers intend to make this semantic distinction. They
11
The Ekman and Friesen (1978) Facial Action Coding System has also been used by other researchers investigating
non-manuals in signed languages (e.g., Dachkovsky & Sandler 2009). We chose not to use this system because it would have required a significant additional investment in resources for annotators to learn it. The lack of any real change in identifying form/meaning pairings using the system used for BSL instead of the original gloss-based system leads us to believe that the even greater detail of the facial coding system would also not have altered the outcome. 12
Anecdotally, the first author can report that most non-native signers have great difficulty reading fingerspelling without mouthing support, so it is not surprising that fingerspelling attracts such high rates of mouthing—the vast majority of signers in the Auslan-using community are not native signers. It would thus be habitual for even a native signer to mouth, as our data show. 13
A correction is in order. Auslan SPOUSE is no more ambiguous for husband or wife than Chinese TA is for he or she or
it. The languages simply do not make a gender distinction at the lexical level for these concepts. SPOUSE unambiguously means ‘person to whom one is married’ and TA means ‘a singular person/entity which is neither first or second person referent’. They are only ‘ambiguous’ for someone in whose language and culture a gender or animacy distinction is obligatorily made.
37
certainly never mouth ‘spouse’. (In fact, some Auslan signers would not know what this English word
means.) However, sampling the much larger Auslan Corpus reveals that, of 37 tokens of SPOUSE, 23 are
mouthed (as ‘wife’, ‘husband’, ‘marry/ied’) and 14 have no mouthing (in all of these senses).14
The fact that
fully one third of SPOUSE tokens have no mouthing suggests that the use of mouthing is not obligatory and
not specified in the signer’s mental lexicon. In the dataset, gender or grammatical function is often simply
inferred from context without any formal (including mouthed) marker: e.g., POSS1SG SPOUSE (unmouthed,
signed by a man) means my wife; or PRO1SG SPOUSE PRO3SG (unmouthed) means I married him/her.
Nonetheless, basic core culturally shared distinctions are made and expected. For Auslan signs, this does
mean a default mouthing of ‘wife’, ‘husband’, or even, ‘marry’ when signing SPOUSE, even if it is not
obligatory. Their interlocutors, who are often second language learners of Auslan and speakers of English as
a first language, give and demand clarification.
A similar default coupling occurs with vocabulary items that name basic level cognitive categories that
one can expect to recur and overlap cross-linguistically and cross-culturally, e.g., tree, flower, boy, girl, dog,
cat, mother, father, man, woman and so on (Taylor 1991). The fact that signers and speakers share
essentially the same culture, reinforces the likelihood of overlap and convergence. The fact that Auslan and
English are in different modalities means that double coding (mouthed in English while manually signed in
Auslan) can occur without any expressive or receptive conflict. Indicative mouthing rates for these types of
concepts from the data set include: TREE (87%), FLOWER (78%), BOY (77%), GIRL (85%), DOG (95%), CAT
(72%), MOTHER (84%), FATHER (100%), MAN (78%), WOMAN (80%), LADY (91%).
Moreover, despite the fact that many signs are associated with a standard mouthing, as shown earlier,
across the lexicon the relationship of manual signs to mouthing is simply not one-to-one: it is one to many.
Most signs can co-occur with different mouthings (and sometimes mouth gestures) that specify, limit, or
extend the core meaning of the sign (Johnston & Schembri 1999; Bank et al. 2011).
The reported findings thus confirm that mouthing is not obligatory in the sense that they are not part of the
lexical specification of any sign, a position long-held by Auslan researchers (Johnston & Schembri 1999)
and other researchers (e.g., Ebbinghaus & Hessmann 2001). Rather, mouthings represent the parallel or
simultaneous expression of elements from a second language, and thus the phenomenon is best described as
14
The fact that an Auslan sign may function as a noun or a verb is unremarkable. Remember, the ID gloss does not capture this distinction.
38
code-blending (Emmorey, Borinstein, & Thompson 2005; van den Bogaerde & Baker 2005; Bank et al.
submitted). Data on slips of the hand and tongue by deaf and hearing BSL signers, suggesting that the
mouthing and manual components have separate mental representations, supports this position (Vinson et al.
2010).
Mouthings are ‘obligatory’ only in the sense that if an Auslan signer wishes to make the semantic
distinction afforded by the mouthing in a situation where context does not render this inferable—and
continue to use just the intended manual sign—then they must use that mouthing. Otherwise, they must
choose another way of incorporating English (e.g., by an interpolated fingerspelling) or instead use
circumlocution in Auslan.
Researchers have suggested that mouthing is primarily or solely a language contact phenomenon or that
formal or oralist education has had an impact on mouthing rates (Hohenberger & Happ 2001; Nadolske &
Rosenstock 2007; Bank et al. 2011, submitted). With respect to the former, though we do not have data on
our corpus participants signing to non-native or hearing signers for comparison, our data are taken from sets
of two native (or near-native) signers (and each knew the other was a fluent signer) with a third native signer
researcher (known to be native to the other two). Our data do show that mouthing is present in large degree
in most texts, so immediate contact in the context of utterance with non-native or hearing signers cannot be
driving this behaviour.
With respect to the impact of formal education for deaf children, reports of deaf signers from communities
who have no formal education and are illiterate show that they use spoken language related mouthings in
their sign production (Nyst 2007; Fontana 2008). For example, Fontana observes that for home signers in
Okinawa and Italian SL signers in Sicily, speech is part of the communicative economy of deaf individuals
and deaf communities everywhere. In the life experience of all deaf people, there will be inevitable exposure
to some signs and gestures that are invariably accompanied by the same mouthing when performed by
hearing people (signers or non-signers). Deaf people cannot help but observe this: observing and
internalizing mouth actions (including mouthing) is an inevitable part of being a deaf person in a hearing
world.
39
Consequently, in Auslan, as in NGT (Bank et al. 2013) and presumably many other signed languages,
mouthing is ubiquitous.15
As Bank et al. have also remarked, this has implications for the independence of
signed from spoken languages in developed urban communities with long histories of deaf education and
widespread basic literacy. We would agree, at minimum, that monolingual speakers of these languages do
not, in fact, exist. All signers display some degree of bilingualism and thus language contact.
Form/meaning pairings in Auslan The majority of mouth gestures (71.4%) were categorized as W-type
(Figure 4) and are indexic expressions, or mimetic and gestural enactments of one kind or another. The small
number of 4-type mouth gestures are also mimetic. Of the A-type mouth gestures, we have determined that
fully two in three appear to fulfil a prosodic function, with little evidence of language specific codification.
Of the remaining ‘tradtional’ non-prosodic (adverbial) A-types, we consider some 45% to be iconic. We also
identified the use of A-types within W-types (during constructed action) and this may be relevant to the
question of conventionalization. It would appear that many adverbial A-type mouth gestures represent
expressive or iconic/mimetic mouth enactments that can also be used with some signs when no overt
constructed action is actually being performed, rather than being examples in which constructed action has
‘co-opted’ adverbial A-types. It is not clear to what we extent we can consider the above mouth gesture types
as specific to any signed language, and it is likely that many form part of a more general semiotic system.
The remaining non-iconic A-types vary in form and effect. It was difficult to identify recurring form
groups and the meanings they conveyed except in the most general of terms. They seemed to have more
subtle, global meanings, to be read in combination with other aspects of the delivery, rather than a specific,
defined meaning.
There has only ever been one good potential candidate in this category: TH. It has long been identified as
the archetypical A-type modifying mouth gesture ‘morpheme’ meaning something like ‘carelessly’ (or ‘with
ease’ which is a suggested second distinctive meaning proposed for BSL, see Lewin & Schembri 2011).
However, the use of this mouth-centred facial gesture does not appear to be specific to signed languages.
Studies, some that predate the work of Liddell (1980), on expressive and communicative human facial
15
Kata Kolok, the signed language of a Balinese village with multi-generational deafness, has been reported to have ‘virtually no mouthings’ (Vos & Zeshan 2012). We too have individuals in our dataset who have extremely low levels of
mouthings (but they still use a considerable number of mouth gestures), so this is not surprising. Lack of formal education and/or an integrated fingerspelling system together with cultural practices that may lead hearing interlocutors to believe that speaking or even mouthing to deaf people is pointless could well have this effect on a community-wide basis.
40
expressions have documented the meanings and functions of tongue protrusion (TH) in great detail (e.g.,
Smith, Chase, & Lieblich 1974; Rozin & Cohen 2013). Few signed language researchers, apart from Fontana
(2008), appear to be aware of this work on humans and primates.
Data from human children and adults show that tongue protrusion is common and is used to index and
communicate distaste, displeasure or disgust; attention or care; recklessness or carelessness;
unencumberedness or ease; and distain or rejection. Some of these meanings clearly overlap the values of
‘carelessly’ or ‘with ease’ that are attributed to this mouth gesture in signed languages. However, the
apparent specificity of these meanings to these signed languages has much to do with how one describes the
meanings of various facial expression (Elliot & Jacobs 2013). The research of Smith et al. (1974), for
example, shows that there are really only two underlying functions of tongue protrusion (a) indexing
distaste/displeasure/disgust and (b) expressing ‘social exclusion’ (i.e., a desire to be uninterrupted or a
rejection of social engagement), with the former probably lying at the basis of the latter (one can readily
appreciate how physical rejection could extend metaphorically to social rejection). When an appropriate
general label is given to the latter (which Smith et al. call ‘social exclusion’), one can see how its semantic
force would vary according to what co-occurs at the time of the tongue protrusion (i.e., the context) and the
positive or negative overall facial expression it is made with. For example, signing DRIVE + negative face +
TH = recklessly, carelessly, i.e., ‘drive in such a way as to reject the concerns or attitudes of others, namely
carelessly’; DRIVE + positive or neutral face + TH = carefree, easily, i.e., ‘drive in such a way as to reject the
concerns or attitudes of others as unnecessary because it is not difficult and you are competent, namely with
ease’; CUT + neutral face + squint + TH = cut with one’s full attention, i.e., keeping (social) interruptions at
bay, namely carefully, attentively or precisely). Signers are able to produce a specific conventional meaning
with a manual sign while they produce a mouth gesture. This provides signers with greater control on the
context in which inferences using the ‘social exclusion’ gesture TH could be constrained. Speakers, on the
other hand, need their mouths to speak so do not have the opportunity to do this.16
The conventionality lies not in the TH mouth gesture but in the conventional signs that with which they co-
occur (and the wider signed context). If one does not know a conventional lexical manual sign meaning
‘drive’ is being produced (driving is an act one would normally do with care and attention), one lacks the
16
Indeed, it is this very phenomenon that prompts Sandler (2009, p. 241) to suggest that the site of ‘co-language’ gesture is inverted in signed languages (“Speakers gesture with their hands, and signers gesture with their mouths”). However, we do not think the division of labor is as neat or as categorical as this suggests (see Conclusion).
41
intended and assumed inferential context. The fact that a similar mouth gesture has been identified in quite a
few signed languages leaves little doubt it is not language-specific. Finally, the fact that no human cultures
so far investigated appear to lack it suggests it is not specific to language, as such.
This leaves very few other mouth gestures that may be considered candidates as conventional language-
specific mouth gestures. The E-type mouth gesture PAH, is one possible example. First of all, like the
example of TH above, one needs to discount one meaning associated with it as overly specific and derived
from an association with a frequently, but not exclusively, co-occurring lexical sign: FINISH (or alternatively
glosses as SUCCESS or FINALLY in both Auslan and BSL). On the other hand, one may argue that it has begun
to be interpreted as a semantic component (‘suddenly’) which can be added to any sign, regardless of
whether or not it echoes that sign’s movement parameter. For example, PAH occurs with SLEEP to mean ‘fell
swiftly into a deep sleep’ which is not a standard pairing (the mouth gesture opens, where as the handshape
closes). It also occurs with depicting signs which do not have a standard form and thus no standard mouth
gesture. (They are productive signs with both conventional components, like handshape, and nonce
components, like placement and movement.) There may even be no formational parameter which is ‘echoed’
in the PAH.
A second E-type candidate may be WOOF because in the study dataset it always occurred with
AUTOMATIC. However, there are two reasons why it clearly would be incorrect to attribute “automaticity” to
the meaning of WOOF. First, across the entire Auslan corpus there are attested examples of AUTOMATIC also
co-occurring with FAA, ALAM, and FEE, all with the same effect as WOOF, with the mouth gesture type TRILL,
and the mouthing ‘own’ (probably an abbreviation for “on its own”). Second, it also occurs with no
mouthing or mouth gesture.
So though the strongest claims to linguistic status have been made for some E- and A-types of mouth
gestures, it would appear that their language-specific conventionalization is, at best, incipient. Moreover, as a
group they only occur in small numbers and may be considered to be a marginal phenomenon.
Overall, the data support earlier suggestions that few Auslan mouth gestures have very clearly specifiable
language-specific forms and are obligatory in any meaningful sense (Johnston & Schembri 2007). Rather
than being a part of the phonological specification of some signs or separable formal morphemes added to
signs and phrases, they are on the whole contextual symbolic units (gestures of one kind or another) and/or
are prosodic.
42
Conclusion
Mouth actions, other facial expressions, head and body movements, and other aspects of sign articulation
(e.g., speed and stress) all contribute in various complex ways to meaning creation in Auslan (Johnston &
Schembri 2007; see also references found at footnotes 1 and 2). However, positing a conventional codified
system for mouth actions as part of this dynamic appears not to be supported by the usage evidence, at least
for Auslan.
The strongest candidates for conventional status were the so-called lexically specified mouth actions
(mouthings and echo phonology). But in both cases, no specifiable mouth action appears to obligatorily co-
occur with particular signs, and exceptions and variation abound. Mouthings appear to be examples of
language blending, even if constant and intimate language contact does mean many signs have a standard
paired mouthing. Mouth gestures do add meaning to co-articulated signs but these meanings are relatively
general and the actual mouth forms employed vary from person to person. Overall, we have seen that there
are often no mouth actions at all in environments in the data were it would be assumed or predicted to occur
had the mouth action the linguistic role attributed, e.g., mouthings with fingerspelling or with lexical
specification, or mouth gestures of a certain form with classes of signs with particular morpho-phonological
characteristics. These facts do not suggest highly codified usage.
Mouth gestures also appear to have parallels in face-to-face spoken language production, although there
has been virtually no research on non-manual gesture use in spoken language. Although infrequent (because
speakers are most of the time busy speaking so the mouth is not available to gesture), observation of day-to-
day spoken language interactions suggests a hearing person’s manual gesture can be accompanied by a
mouth gesture which echoes the dynamics of the former, e.g., a fist opening to five handshape with an
opening mouth (from a plosive-like to an open vowel-like syllable) to express rapidly throwing something,
something suddenly opening or exploding, a spitting or expelling action, or something splattering on a
surface.
In this context, it would seem that the definition often given of mouth gestures as mouth actions that are
unlike mouthing in that they ‘do not derive from spoken language’ (Boyes Braem & Sutton-Spence 2001)—
as if they arose spontaneously within signed languages and were sign language dedicated—needs
qualification. Most mouth gestures that occur in signed languages appear to be part of the repertoire of facial
expressions used by members of the wider culture in which deaf signers find themselves. Of course, some
43
mouth gestures may be negatively sanctioned in some cultures and, thus, be suppressed and rarely produced,
yet they would remain comprehensible to anyone in such a culture given the appropriate context.
So mouth gestures also are often sourced from spoken language insofar as speakers constitute the vast
majority of any human community. Everyone, signers included, imitate the faces of the people they interact
with on a daily and ‘face-to-face’ level. It is true that many mouth actions are actually the articulation of
spoken words (mouthings as far as a deaf person is concerned), but many are not. Most of the data do not
suggest that all mouth gestures are autonomous to signed languages or that they have their own unique
meaning and function in these languages.
We would conclude that mouth actions in Auslan are thus significantly less conventional and more
gestural than has been proposed in the literature for a number of other signed languages. We believe that
other signed languages may not be very different, but certainty in this regard requires analysing much larger
and representative datasets of naturalistic data then has hitherto been the case.
We believe mouth actions exist along a continuum of indexicality, iconicity and conventionality (Figure
10). Highly conventional (i.e., linguistic) mouth gestures, should they exist in some signed languages, would
only account for a small proportion of all mouth actions in any signed language. Mouthings are linguistic in
the sense that they are instances of code blending.
Figure 10 A continuum of conventionalization in mouth actions
Mouth actions work as part of signed languages which we believe, like all human languages, are best
described as heterogeneous semiotic systems (Schembri 2001; de Beuzeville, Johnston, & Schembri 2009;
Johnston 2013). Mouth actions are additions that support other facets of the utterance to create a synthesized
meaning in context. These mouth actions, both the mouth gestures and the mouthings, add meaning to the
primary conventional semiotic units of the linguistic system, the symbolic constructions of the language (the
individual signs and multi-sign strings). In this sense, mouth actions as a whole perform a role not unlike that
44
often attributed to co-speech gesture: adding meaning which is formally un-stated or under-stated in the
wording (i.e., the linguistic coding). However, it would be a mistake to equate mouth actions with co-speech
gesture if by this we imply this is the primary or sole locus of ‘gesturing’ within signed languages (cf.
Sandler 2009). We do not believe that it is only on the mouth that gesture-in-sign resides (‘co-sign gesture’
as it were) because gestural elements can also be found in other levels of the lexico-grammar of these
languages (Liddell 2003; Wilcox 2004; de Beuzeville et al. 2009; Johnston & Schembri 2010; Janzen 2012).
They are not only manifested in mouth actions.
References
Aarons, D., Bahan, B., Kegl, J. A., & Neidle, C. (1992). Clausal Structure and a Tier for Grammatical
Marking in American Sign Language. Nordic Journal of Linguistics, 15, 103-142.
Anderson, D. E., & Reilly, J. S. (1998). PAH! The acqusition of non-manual adverbials in ASL. Sign
Language & Linguistics, 1(2), 115-142.
Baker, C., & Padden, C. (1978). Focusing on the Non-manual Components of American Sign Language. In
P. Siple (Ed.), Undertanding Language through Sign Language Research (pp. 27-58). New York:
Academic Press.
Baker-Shenk, C. (1983). A microanalysis of the non-manual components of questions in American Sign
Language. doctoral dissertation, University of California, Berkeley, CA.
Bank, R., Crasborn, O., & van Hout, R. (2011). Variation in mouth actions with manual signs in Sign
Language of the Netherlands (NGT). Sign Language & Linguistics, 14(2), 248-270.
Bank, R., Crasborn, O., & van Hout, R. (2013). Alignment of two languages: The spreading of mouthings in
Sign Language of the Netherlands. International Journal of Bilingualism, 0(0), 1-16.
Bank, R., Crasborn, O., & van Hout, R. (submitted). The prominence of spoken language elements in a sign
language.
Bergman, B. (1984). Non-manual components of signed language: some sentence types in Swedish Sign
Language. In F. Lonke, P. Boyes Braem & Y. Lebrun (Eds.), Recent research on European sign
languages (pp. 49-59). Lisse: Swets & Zeitlinger.
Boyes Braem, P. (2001). Functions of the mouthings in the signing of deaf early and late learners of Swiss
German Sign Language (DSGS). In P. Boyes Braem & R. Sutton-Spence (Eds.), The hands are the
head of the mouth: The mouth as articulator in sign languages (pp. 99-131). Hamburg: Signum
Press.
Boyes Braem, P., & Sutton-Spence, R. (Eds.). (2001). The hands are the head of the mouth: The mouth as
articulator in sign languages. Hamburg, Germany: Signum Press.
Brentari, D., & Padden, C. (2001). Native and foreign vocabulary in American Sign Language: A Lexicon
With Multiple Origins. In D. Brentari (Ed.), Foreign Vocabulary in Signed Languages: A Cross-
Linguistic Investigation of Word Formation (pp. 87-120). Mahwah, NJ: Lawrence Erlbaum
Associates.
Cormier, K., Schembri, A., & Tyrone, M. (2008). One hand or two? Nativisation of fingerspelling in ASL
and BANZSL. Sign Language & Linguistics, 11(1), 3–44.
Crasborn, O., van der Kooij, E., Waters, D., Woll, B., & Mesch, J. (2008). Frequency distribution and
spreading behavior of different types of mouth actions in three sign languages. Sign Language &
Linguistics, 11(1), 45-67.
Cuxac, C. (1999). French Sign Language: Proposition of a Structural Explanation by Iconicity. In A.
Braffort, R. Gherbi, S. Gibet, D. Teil & J. Richardson (Eds.), Gesture-Based Communication in
Human-Computer Interaction: International Gesture Workshop, GW'99 Gif-sur-Yvette, France,
March 17-19, 199 Proceedings. Berlin: Springer-Verlag.
Cuxac, C. (2000). La Langue des Signes Française (LSF): Les voies de l'iconicité. Paris: Ophrys.
Dachkovsky, S., & Sandler, W. (2009). Visual Intonation in the Prosody of a Sign Language. Language &
Speech, 52(2-3), 287-314.
45
de Beuzeville, L., Johnston, T., & Schembri, A. (2009). The Use of Space with Indicating Verbs in Auslan:
A corpus based investigation. Sign Language & Linguistics, 12(1), 53-82.
Ebbinghaus, H., & Hessmann, J. (1990). German words in German Sign Language: Theoretical
Considerations Prompted by an Empirical Finding. In S. Prillwitz & T. Vollhaber (Eds.), Current
Trends in European Sign Language Research: Proceedings of the 3rd European Congress on Sign
Language Research Hamburg July 26-29, 1989 (pp. 97-114). Hamburg: Signum Verlag.
Ebbinghaus, H., & Hessmann, J. (2001). Sign language as multidimensional communication: Why manual
signs, mouthings, and mouth gestures are three different things. In P. Boyes Braem & R. Sutton-
Spence (Eds.), The hands are the head of the mouth. The mouth as articulator in sign languages (pp.
133-151). Hamburg: Signum Press. Efron, D. (1972). Gesture, Race and Culture. The Hague: Mouton & Co
Elliot, E. A., & Jacobs, A. M. (2013). Facial expressions, emotions, and sign languages. Frontiers in
Psychology, 4, 1-4.
Emmorey, K., Borinstein, H. B., & Thompson, H. (2005). Bimodal Bilingualism: Code-blending between
Spoken English and American Sign Language. In J. Cohen, K. T. McAlister, K. Rolstad & J.
MacSwan (Eds.), ISB4: Proceedings of the 4th International Symposium on Bilingualism.
Sommerville, MA: Cascadilla Press.
Engberg-Pedersen, E. (1993). Space in Danish Sign Language: The semantics and morphosyntax of the use
of space in a visual language. Hamburg: Signum Press.
Fenlon, J., Schembri, A., Rentelis, R., Vinson, D., & Cormier, K. (2014). Using conversational data to
determine lexical frequency in British Sign Language: The influence of text type. Lingua, 143(187-
202).
Fontana, S. (2008). Mouth actions as gesture in sign language. Gesture, 8(1), 104-123.
Hohenberger, A., & Happ, D. (2001). The linguistic primary of signs and mouth gestures over mouthing:
Evidence from language proudvtion in German Sign Language (DGS). In P. Boyes Braem & R.
Sutton-Spence (Eds.), The hands are the head of the mouth: The mouth as articulator in sign
languages. Hamburg: Signum Press.
Janzen, T. (2012). Lexicalization and grammaticalization. In R. Pfau, M. Steinbach & B. Woll (Eds.), Sign
Languages: An International Handbook (pp. 816-840). Berlin: De Gruyter.
Johnson, D. E. (2009). Getting off the Goldvarb standard: Introducing Rbrul for mixed-effect variable rule
analysis. Language and Linguistics Compass, 3, 359–383.
Johnston, T. (1992). The Realization of the Linguistic Metafunctions in a Sign Language. Language
Sciences, 14(4), 317-353.
Johnston, T. (2010). From archive to corpus: transcription and annotation in the creation of signed language
corpora. International Journal of Corpus Linguistics, 15(1), 104-129.
Johnston, T. (2012). Lexical Frequency in Sign Languages. Journal of Deaf Studies and Deaf Education,
17(2), 163-193.
Johnston, T. (2013). Towards a comparative semiotics of pointing actions in signed and spoken languages.
Gesture, 13(2), 109-142.
Johnston, T., & Ferrara, L. (2012). Lexicalization in signed languages: when is an idiom not an idiom?
Proceedings of the 3rd UK Cognitive Linguistics Conference, University of Hertfordshire, 6-8 July
2010, 1, http://www.uk‐cla.org.uk/proceedings.
Johnston, T., & Schembri, A. (1999). On defining lexeme in a sign language. Sign Language & Linguistics,
2(2), 115-185.
Johnston, T., & Schembri, A. (2007). Australian Sign Language (Auslan): An introduction to sign language
linguistics. Cambridge: Cambridge University Press.
Johnston, T., & Schembri, A. (2010). Variation, lexicalization and grammaticalization in signed languages.
In B. Garcia & M. Derycke (Eds.), Sourds et langue des signs: Normes et variation. Langage et
société 131, 5-17.
Kendon, A. (1980). Gesticulation and speech: Two aspects of the process of utterance. In M. R. Key (Ed.),
Nonverbal communication and language (pp. 207–277). The Hague: Mouton and Co
Kendon, A. (2004). Gesture: Visible Action as Utterance. Cambridge: Cambridge University Press.
Kendon, A. (2008). Some reflections on the relationship between 'gesture' and 'sign'. Gesture, 8(3), 348-366.
Krahmer, E., & Swerts, M. (2009). Audiovisual Prosody—Introduction to the Special Issue. Language and
Speech, 52(2/3), 129-133.
Lawson, L. (1983). Multi Channel Signs in British Sign Language. In J. G. Kyle & B. Woll (Eds.), Language
in Sign: an international perspective on sign language (pp. 97-105). London: Croom Helm.
46
Lewin, D., & Schembri, A. (2011). Mouth gestures in British Sign Language (BSL): A case study of tongue
protrusion in BSL narratives. Sign Language and Linguistics, 14(1), 94-114.
Liddell, S. K. (1978). Nonmanual Signals and Relative Clauses in American Sign Language. In P. Siple
(Ed.), Understanding Language Through Sign Language Research (pp. 59-90). New York:
Academic Press.
Liddell, S. K. (1980). American Sign Language Syntax. The Hague: Mouton.
Liddell, S. K. (2003). Grammar, Gesture, and Meaning in American Sign Language. Cambridge: Cambridge
University Press.
Lucas, C. (Ed.). (2006). Multilingualism and Sign Languages: From the Great Plains to Australia (Vol. 12).
Washington, DC: Gallaudet University Press.
McIntire, M. L., & Reilly, J. S. (1988). Non-manual behaviours in L1 and L2 learners of American Sign
Language. Sign Language Studies, 61, 351-376.
McKee, D., & Kennedy, G. (2006). The Distribution of Signs in New Zealand Sign Language. Sign
Language Studies, 6(4), 372-390.
McNeill, D. (1992). Hand and mind: What gestures reveal about thought. Chicago: University of Chicago
Press.
McNeill, D. (2012). How Language Began: Gesture and Speech in Human Evolution. Chicago: Chicago
University Press.
McNeill, D. (Ed.). (2000). Language and gesture. Cambridge: Cambridge University Press.
Meir, I., Padden, C., Aronoff, M., & Sandler, W. (2013). Competing iconicities in the structure of languages.
Cognitive Linguistics, 24(2), 309-343.
Militzer, S. (2013). The correlation between word class and mouth actions in Irish Sign Language. Paper
presented at the TISLR, London, UK.
Morford, J., & MacFarlane, J. (2003). Frequency Characteristics of American Sign Language. Sign
Language Studies, 3(2), 213-225.
Nadolske, M., & Rosenstock, R. (2007). Occurrence of mouthings in American Sign Language: A
preliminary study. In P. Perniss, R. Pfau & M. Steinbach (Eds.), Visible Variation: Comparative
studies on sign language structure (pp. 35-61). Berlin: Mouton de Gruyter.
Neidle, C., Kegl, J. A., Maclaughlin, D., Bahan, B., & Lee, R. (2000). The Syntax of American Sign
Language: Functional Categories and Hierarchical Structure. Cambridge, MA: MIT Press.
Nespor, M., & Sandler, W. (1999). Prosody in Israeli Sign Language. Language and Speech, 42(2-3), 143-
176.
Newport, E. L., & Supalla, S. J. (2000). Sign Language Research at the Millennium. In K. Emmorey & H.
Lane (Eds.), The signs of language revisited: An anthology to honor Ursula Bellugi and Edward
Klima (pp. 103-114). Mahwah, NJ: Lawrence Erlbaum Associates.
Nyst, V. (2007). A Descriptive Analysis of Adamorobe Sign Language (Ghana). Utrecht, The Netherlands:
LOT.
Padden, C. (1980). The Deaf Community and the Culture of Deaf People. In C. Baker & R. Battison (Eds.),
Sign Language and the Deaf Community: Essays in Honor of William C. Stokoe (pp. 89-104). Silver
Spring, MD: The National Association of the Deaf.
Perniss, P., Thompson, R., & Vigliocco, G. (2010). Iconicity as a general property of language: Evidence
from spoken and signed languages. Frontiers in Psychology, 1.
Pfau, R., & Quer, J. (2010). Nonmanuals: their grammatical and prosodic roles. In D. Brentari (Ed.), Sign
Languages (pp. 381-402). Cambridge: Cambridge University Press.
Pizzuto, E. (2003). Review of Boyes Braem, P., & Sutton-Spence, R. (Eds.). (2001). The hands are the head
of the mouth: The mouth as articulator in sign languages. Hamburg, Germany: Signum Press. Sign
Language and Linguistics, 6(2), 300-305.
Quinto-Pozos, D. (Ed.). (2007). Sign Languages in Contact. Washington, D.C.: Gallaudet University Press.
Rand, D., & Sankoff, D. (1991). GoldVarb 2.1: A Variable Rule Application for Macintosh. Montreal:
Centre de recherches mathematiques.
Rozin, P., & Cohen, A. B. (2013). High Frequency of Facial Expressions Corresponding to Confusion,
Concentration, and Worry in an Analysis of Naturally Occuring Facial Expressions of Americans.
Emotion, 3(1), 68-75.
Sandler, W. (1999a). The medium and the message: Prosodic interpretation of linguistic content in Israeli
Sign Language. Sign Language & Linguistics, 2, 187-215.
Sandler, W. (1999b). Prosody in Two Natural Language Modalities. Language and Speech, 42(2-3), 127-
142.
47
Sandler, W. (1999c). Cliticization and prosodic words in a sign language. In T. A. Hall & U. Kleinhenz
(Eds.), Studies on the phonological word (pp. 223-254). Amsterdam: John Benjamins.
Sandler, W. (2009). Symbiotic symbolization by hand and mouth in sign language. Semiotica, 174, 241-275.
Sandler, W. (2012). Dedicated gestures and the emergence of sign languages. Gesture, 12(3), 265-307.
Schembri, A. (2001). Issues in the analysis of polycomponential verbs in Australian Sign Language (Auslan).
doctoral dissertation, University of Sydney, Sydney.
Schermer, T. (1985). Analysis of natural discourse of deaf adults in the Netherlands: Observations on Dutch
Sign Language. In W. C. Stokoe & V. Volterra (Eds.), Prodeedings of the IIIrd International Sign
Language Research Symposium, Rome, June 22-26, 1983. Silver Spring, MD: Linstok Press.
Schermer, T. (1990). In search of a language: Influences from spoken Dutch on Sign Language of the
Netherlands. PhD Dissertation, University of Amsterdam. Delft: Eburon.
Schermer, T. (2001). The Role of Mouthings in Sign Language of the Netherlands: Some Implications for
the Proudction f Sign language Dictionaries. In P. Boyes Braem & R. Sutton-Spence (Eds.), The
hands are the head of the mouth: The mouth as articulator in sign languages (pp. 66-87). Hamburg:
Signum Press.
Schroeder, O. (1985). A problem in phonological description. In W. C. Stokoe & V. Volterra (Eds.), SLR '83
(pp. 194-201). Silver Spring, MD: Linstok Press.
Siu Wai-yan, R. (2007). Mouth Gestures in Hong Kong Sign Language (HKSL): A Preliminary Study.
Unpublished Master of Arts dissertation, Department of Linguistics, Chinese University of Hong
Kong.
Smith, W. J., Chase, J., & Lieblich, A. K. (1974). Tongue showing: A facial display of humans and other
primate species. Semiotica, 11, 201-246.
Sutton-Spence, R., & Day, L. (2001). Mouthings and mouth gestures in British Sign Language (BSL). In P.
Boyes Braem & R. Sutton-Spence (Eds.), The hands are the head of the mouth: The mouth as
articulator in sign languages (pp. 66-87). Hamburg: Signum Press.
Sutton-Spence, R., & Woll, B. (1999). The linguistics of British Sign Language: an introduction. Cambridge,
U.K.: Cambridge University Press.
Tang, G. (2006). Questions and Negations in Hong Kong Sign Language. In U. Zeshan (Ed.), Interrogative
and Negative Constructions in Sign Languages (pp. 198-224). Nijmegen: Ishara Press.
Tannen, D. (1986). Introducing constructed dialogue in Greek and American conversational and literacy
narratives. In F. Coulmas (Ed.), Reported speech across languages (pp. 311–332). The Hague:
Mouton.
Taub, S. (2001). Language from the Body: Iconicity and Metaphor in American Sign Language. Cambridge:
Cambridge University Press.
Taylor, J. R. (1991). Linguistic Categorization: Prototypes in linguistic theory. Oxford: Clarendon Press.
Torigoe, T., & Wataru, T. (2002). A descriptive analysis of pointing and oral movements in a Home Sign
System. Sign Language Studies, 2(3), 281-295.
van den Bogaerde, B., & Baker, A. (2005). Code mixing in mother-child interaction in deaf families. Sign
Language & Linguistics, 8(1/2), 151-174.
Van Herreweghe, M., & Vermeerbergen, M. (Eds.). (2004). To the Lexicon and Beyond: Sociolinguistics in
European Deaf Communities. Washington, D.C.: Gallaudet University Press.
Vinson, D., Thompson, R., Skinner, R., Fox, N., & Vigliocco, G. (2010). The Hands and Mouth Do Not
Always Slip Together in British Sign Language: Dissociating Articulatory Channels in the Lexicon.
Psychological Science, 21(8), 1158-1167.
Vogt-Svendsen, M. (1983). Lip Movements in Norwegian Sign Language. In J. Kyle & B. Woll (Eds.),
Language in Sign: an international perspective on sign language (pp. 85-96). London: Croom Helm.
Vogt-Svendsen, M. (2001). A comparison of mouth gestures and mouthings in Norwegian Sign Language
(NSL). In P. Boyes Braem & R. Sutton-Spence (Eds.), The hands are the head of the mouth: The
mouth as articulator in sign languages (pp.??). Hamburg: Signum Press.
Vos, C. d., Kooij, E. v. d., & Crasborn, O. (2009). Mixed Signals: Combining Linguistic and Affective
Fuctions of Eyebrows in Questions in Sign Language of the Netherlands. Language and Speech,
52(2/3), 315-339.
Vos, C. d., & Zeshan, U. (2012). Demographic, sociocultural and linguistic variation across rural signing
communities. In U. Zeshan & C. d. Vos (Eds.), Sign languages in village communities:
Anthropological and lingistic insights (pp. 2-23). Boston/Berlin: Walter de Gruyter.
Wilbur, R. B. (1999). Stress in ASL: Empirical evidence and linguistic issues. Language and Speech, 42,
229-250.
48
Wilbur, R. B. (2009). Effects of Varying Rate of Signing on ASL Manual Signs and Nonmanual Markers.
Language and Speech, 52(2/3 ), 245-285.
Wilbur, R. B., & Patschke, C. G. (1999). Syntactic correlates of brow raise in ASL. Sign Language and
Linguistics, 2(1), 3-41.
Wilcox, S. (2004). Gesture and language: Cross-linguistic and historical data from signed languages.
Gesture, 4(1), 43-73.
Woll, B. (2001). The sign that dares to speak its name: Echo phonology in British Sign Language (BSL). In
P. Boyes Braem & R. Sutton-Spence (Eds.), The hands are the head of the mouth: The mouth as
articulator in sign languages. Hamburg: Signum Press.
Woll, B., Sutton-Spence, R., & Elton, F. (2001). Multilingualism: The global approach to sign languages. In
C. Lucas (Ed.), The sociolinguistics of sign languages (pp. 8-32). Cambridge, U.K: Cambridge
University Press.
Zeshan, U. (Ed.). (2006). Interrogative and Negative Constructions in Sign Languages. Nijmegen: Ishara
Press.
Appendix
Mouth gesture form codes and glosses used in this study and their equivalents in the Sutton-Spence
and Day (2001) coding schema from which they were adapted.
‘Closed’ mouth gestures ‘Open’ mouth gestures Tongue-related mouth gestures
Auslan study codes
SS&Day Auslan study glosses
Auslan study codes
SS&Day Auslan study glosses
Auslan study codes
SS&Day Auslan study gloss
CN1 CN14 LIP-CURL ON1 OL12 OPEN OT1TL OL18 TONGUE
CN2 CN16 WIDE ON2 OL13 OPEN OT1TM OL19 TONGUE
CN3 CN15 n/a ON3 OL13 OPEN OT1TR OL20 TONGUE
CN4 CN17 DOWN ON4 n/a DOWN OT1BL OL18 TONGUE
CN5 CN18? LIPS-PRESSED ON5 n/a n/a OT1BM OL19 TONGUE
CN6 CN18 n/a ON6 OL12 SLIGHTLY OPEN OT1BR OL20 TONGUE
CN7 CN19 TRILL ON7 OL13 WIDE OT1WL OL18 TONGUE
CN8 CN20 BLOW ON OL13 WIDE OT1WM OL19 TONGUE
CN9 CN22 TRILL ON9 n/a DOWN OT1WR OL20 TONGUE
CN10 CN21 TRILL ON10 n/a LIP-CURL OT2 OL15 TONGUE
CN11 CN23 LIPS-OUT ON11 OL14 BOTTOM-LIP-OUT OT3 n/a TONGUE
CN12 CN7 LIPS‐OUT ON12 OL6 SLIGHTLY OPEN OT4TL OL9 TONGUE
CN13 CN8 LIPS-OUT/TRILL ON13 OT7 WIDE OT4TM OL9 TONGUE
CN14 CN9 LIPS-OUT ON14 OL7 WIDE OT4TR OL9 TONGUE
CN15 CN12 TRILL ON15 n/a DOWN OT4BL OL10 TONGUE
CN16 CN13 LIPS-OUT ON16 OL8 LIPS-OUT OT4BM OL10 TONGUE
CN17 CN10 TRILL ON17 n/a BLOW OT4BR OL10 TONGUE
CN18 CN11 TRILL ON18 n/a BLOW OT4WL OL11 TONGUE
CN19 CN2 TONGUE ON19 n/a BLOW OT4WM OL11 TONGUE
CN20 CN3 BOTTOM‐LIP‐OUT OT4WR OL11 TONGUE
CN21 CN4 PRESSED OT5 OT1 TONGUE
CN22 CN5 DOWN OT6 n/a TONGUE
CN23 CN6 PRESSED OT7TL OL1,OT8 TONGUE
OT7TM OL2,OT8 TONGUE
CP1 CN1 PUFF OT7TR OL3,OT8 TONGUE
CP2 CP1 PUFF OT7BL OL4,OT9 TONGUE
CP3 CP3,4,5 PUFF OT7BM OL4,OT9 TONGUE
CP4 CP2 PUFF OT7BR OL4,OT9 TONGUE
CP5 n/a TRILL OT7WL OL5,OT10 TONGUE
CP6 CP7 PUFF/TRILL OT7WM OL5,OT10 TONGUE
CP7 CP6 PUFF/TRILL/BLOW OT7WR OL5,OT10 TONGUE
CP8 CP8 PUFF/BLOW OT8 OT1 TONGUE
OT9 n/a TONGUE