PLIN7309 The linguistics of sign languages

1
PLIN7309 The linguistics of sign languages

• Sign language morphology 2
• Adam Schembri

2
Morphemes on the face 1

• Structure and organisation of sign languages
  filtered through the whole body and face, and
  are not confined to the hands (S L-M, 2006
  60)
• Specific non-manual features occur with content
  and polar questions and delineate prosodic
  constituents equivalent to intonation (e.g.,
  content question brow furrow head tilt
  forward)
• Other non-manual features associated with
  multi-channel signs
• BSL NOT-YET (sh), SPOT-ON (pee), TRUE (hup)
• ASL NOT-YET (ee) vs LATE (th)
• BSL/ASL RECENT vs VERY-RECENTLY (csee)
• Some appear productive
• ASL DRIVE (mm) vs DRIVE (th)

3
Morphemes on the face 2

• These non-manual features typically occur with
  verbs and act as adverbials
• Thus, within the domain of word formation (unlike
  intonational uses of non-manual features)
• They are bound morphemes, as they do not appear
  alone
• They are represent a minimal pairing of form and
  meaning classical definition of a morpheme
• But not clear how they are unanalysable
  phonologically

4
Morphomes in ASL 1

• Aronoff (1994) morphome is a grammatical
  morpheme with more than one function
• Originally distinct morphemes that have merged
  diachronically, resulting in syncretism?
• -s in English
• jogz, billz manz
• walks, hats, shops
• lunchez, churchez, judgeez
• Person/number/tense plural possessive

5
Morphomes in ASL 2

• In ASL, alternating reduplication is a morphome
• Exhaustive inflection characteristic adjective
  derivation
• Three allomorphs alternating, non-alternating or
  one-handed reduplication, depending on
  phonological characteristics of the base
• 1h base alternating reduplication
• GIVEexhaustive CHARACTERISTICALLY-UNDERTANDING
• 2h symmetrical base non-alternating
  reduplication
• SIGNexhaustive CHARACTERISTICALLY-QUIET
• 2h asymmetrical base one-handed reduplication
• REJECTexhaustive CHARACTERISTICALLY-CRUEL

6
Classifier constructions 1

• Classifier constructions can be grouped into
  subcategories based on their semantics and other
  characteristics what are they?
• (1) TWO MAN CL two-animate-entities(fmove-line
  c) fTELL-OFFc
• The two men approached me and told me off
• (2) SURPRISE MANY BICYCLE CLVehicle(locdistribu
  tionin-a-row)
• Wow, there were many bicycles lined up in rows
• (3) CUP PRO-1 CL Handle-cylindrical-entity(cmov
  e-arcf)
• I hand the glass over to you
• (4) HOME HAVE TABLE CL Two-dimensional-outline(t
  racekidney-shape)
• At home there is a kidney shaped table

7
Classifier constructions 2

• CCs are a widespread but anomalous subsystem in
  sign language grammars
• Their form, although considered morphologically
  complex, is different from complex lexical signs
• Although they can act as a verb or as whole
  proposition, they not like ordinary verbs or
  sentences either
• The handshape in such signs has been analysed as
  representing different classes (thus
  classifiers) of nominals in combination with
  other elements

8
Classifier constructions 3

• There are a number of anomalous characteristics
  of CCs
• (1) Verbs? CCs are unlike other verbs because
  entity CCs do not take aspectual inflections
• (2) Each hand can represent a different nominal
  argument, unlike lexical signs
• Phonologically, they are unlike words because
• (3) Structural elements in CCs are not
  meaningless formational elements, but may be
  morphemic, unlike lexical signs
• (4) They may violate the Symmetry Condition as
  each hand may move freely
• (5) They may violate the Dominance Condition, as
  the non-dominant hand may be marked

9
Classifier constructions 4

• Propositional content and prosodic structure are
  different in CCs than from lexical signs as well
• (6) Complex but monosyllabic
• (7) One form may span several intonational
  phrases
• (8) May represent multiple propositions with the
  same form
• S L-M (2006) propose morpho-phonological
  architecture that has three distinct levels
  lexical, postlexical and non-lexical

10
CCs and the architecture of the phonological
component

• Lexical level the addition of morphemes changes
  the phonological structure of signs
• Morphologically modified forms are pronounced
  differently than morphologically simple
  equivalents (e.g., 1-ASK-2, 2-ASK-1)
• These are part of lexical phonology
• Post-lexical level phonological processes that
  occur when lexical items are put together in
  sentences
• Lexical items are pronounced differently due to
  influences from surrounding signs (e.g., location
  assimilation in PRO-1 NAME A-D-A-M)
• These are part of phonological operations at the
  post-lexical level

11
A non-lexical level?

• Sign languages may be unique in possessing a
  non-lexical level
• CCs consist of phonologically underspecified
  (like English -s) morphemes of hand
  configuration, location and movement
• These elements must be in the lexicon, but their
  combination does not result in words they
  combine to form complex predicates
• The clear formal dichotomy between lexical and
  non-lexical subsystems may have another
  explanation, however
• Is it the case that all aspects of CCs are in the
  lexicon, or is there some other explanation?

12
Classifier constructions or depicting verbs?

• Like agreement verbs, a subset of spatial verbs
  (classifier constructions) have also been the
  source of much debate
• In these forms, a conventionalised (?) set of
  handshapes can be moved freely (?) and/or
  positioned in space to describe the way in which
  entities move/are located/are handled or to trace
  their size and shape characteristics

13
Supalla (1982, 1986)

• Claimed all the characteristic aspects of
  classifier constructions are discrete, listable
  and specified in the grammar of each individual
  signed language
• Thus, morphemic status proposed for all movement
  and location units in classifier construction in
  ASL (Supalla, 1982)
• Evidence includes the apparent long developmental
  timetable for classifier constructions, similar
  to morphologically complex constructions in
  spoken languages

14
Liddell (2003)

• Liddell (2003) argued that, like agreement verbs,
  classifier constructions are composed of a
  combination of linguistic and gestural components
• He has proposed that the hand configuration and
  movement component of (some?) classifier
  constructions are lexically specified, while uses
  of location and orientation are gestural

15
Depicting verbs

• HC and MOV lexically specified (for this example
  - MOV may be modified gesturally in other
  examples)
• ORI and POA gestural
• Another example of the fusion of language and
  gesture in the same lexical item

16
Verbs of Motion Production (VMP) Study

• To investigate the claim that depicting verbs in
  signed languages represent fusions of language
  and gesture, comparative data needed
• Research questions
• How similar are depicting verbs of motion
  cross-linguistically?
• How similar are depicting verbs of motion in
  signed languages and gestural representations of
  motion events in non-signers?

17
Materials

• Supalla (1982) devised elicitation materials for
  the study of depicting verbs of motion in ASL
  the VMP Task
• Video stimulus materials suitable for elicitation
  cross-linguistically
• 80 short animated clips showing 10 general
  categories of referents moving in 6 specific ways

18
Target structures

• Materials elicit the following
• Referent categories straight horizontal,
  straight vertical, flat wide, flat narrow,
  circular, cylindrical, animate, vehicle,
  aeroplane and tree referents
• Referent motion linear, turning, random,
  pivoting, jumping/bouncing and falling movements
• Referent spatial relationships initial, middle
  or final contact between referents

19
Method

• The VMP Task used to elicit data from three
  groups of participants (Schembri, Jones
  Burnham, 2005)
• 25 deaf native signers of Auslan (Australian Sign
  Language
• 4 deaf native signers of TSL (Taiwanese Sign
  Language)
• 25 hearing native speakers of Australian English
  with no knowledge of any signed language
• All elicitation sessions videotaped for analysis,
  with 10 of responses independently coded by
  native signer (gt90 agreement)
• The VMP Task has been used in previous studies
  with adult deaf native signers of ASL, deaf
  children using home sign systems, and hearing
  non-signers (Singleton, Morford Goldin-Meadow,
  1993)

20
Results MOV

• Match with ASL MOV targets very high
• Results for MOV Auslan 92, TSL 88, Australian
  gesturers 70
• Differences between Auslan and gesturers
  significant (t(29.49)8.09, plt0.001) but how to
  account for the degree of similarity?
• Compared to ASL (90), home signer (83) and
  American gesturers (95)

21
Results POA

• Match with ASL POA targets very high
• Results for POA Auslan 92, TSL 90, gesture 74
• Differences between Auslan and gesturers
  significant (t(36.19)4.40, plt0.001) but how to
  account for the degree of similarity?
• Compared to ASL (97), home signer (91) and
  American gesturers (73)

22
Results HC

• Results for HC Auslan 57, TSL 44, and
  Australian gesturers 24
• Compared with ASL 84, home signer 46 and
  American gesturers 22
• The difference between Auslan signers and
  Australian gesturers is significantly lower for
  MOV and POA than HC

23
HC categorical use?

• Match with ASL HC targets somewhat irrelevant
• More interesting is the number of distinct HCs in
  responses
• Example responses in the Vehicle category
  Auslan signers used 4, TSL used 3, gesturers 10
• Hearing signers were less categorical, producing
  significantly greater variety of HCs per category
  than Auslan signers (F(1,48)120.06, plt0.001)

24
Others results

• The similarities in the data are striking, and
  not adequately explored or explained in existing
  literature
• Singleton et al. (1993) appear to accept the
  polymorphemic account they claim that standards
  of well-formedness require gradual development
  over time, but this is less true of their data
  (and our data) on MOV and POA than HC
• Goldin-Meadow (2003) used VMP data to argue for
  the emergence of complex morphology in home
  signers, but are her results due to the
  resilience of language, or to resilience of
  language and gesture?

25
Conclusion

• Not clear how to interpret these results, but
  polymorphemic analysis of depicting verbs appears
  in doubt
• The data do suggest significant interpenetration
  and interaction between language and gestural
  uses of the visual-gestural modality
• Provides some evidence for Liddell (2003), but
  more research needed to tease out possible
  differences in gestural and signed responses
  (e.g., segmentation of manner and path in signed
  languages versus co-speech gestures, see Perniss
  Ozyurek, 2007)

26
Other recent evidence aphasia

• Pickell, Klima, Bellugi Hickok (in press)
  report
• 21 unilaterally brain damaged life-long deaf and
  hearing signers (13 LH, 8 RH) given story
  narration task
• LH damaged signers produced significantly more
  errors for lexical signs, RH for depicting verbs
  (I.e., classifier constructions)
• This suggests different patterns of hemispheric
  asymmetry/neural organisation for these sign
  classes
• The requirement to encode analogue gestural
  spatial information increases the involvement of
  the RH

27
Implications

• This model assumes an interaction between
  lexically specified aspects of depicting verbs
  and iconic gesture
• Signed languages thus typologically unique in the
  exact nature of their interaction with gesture
• But such a close relationship between language
  and gesture is not unique to signed languages.

28
Sign, speech and gesture

• Adult language interacts semantically and
  temporally with gestures (McNeill, 1992)
• Complexity and frequency of gestures reflect the
  complexity and fluency of spoken language, in
  both fluent and stuttering speech (Mayberry
  Jaques, 2000)
• The form of iconic gestures can change depending
  on the grammar of the language being spoken (Kita
  Ozyurek, 2003)
• Childrens production of iconic gestures
  dependent on their ability to produce complex
  speech (Ozyurek, 2004 Nicoladis, 2002) iconic
  co-speech gestures may in fact be difficult to
  master
• This research parallels the complex maturational
  time-course identified for depicting verbs in
  signed languages

29
Challenges for the model

• Distinction between morpheme and gesture more
  work needed
• Gesture studies more work needed on the
  structure, use, acquisition, and processing of
  deictic and iconic gestures (both in co-speech
  and no speech conditions)
• Psycholinguistic studies needed how to test
  Liddells claims?
• Clinical and neuro-linguistic studies more
  studies required

30
Conclusion

• The morphological system in signed languages is
  replete with patterns, many of which are found in
  spoken languages as well
• Certain aspects of the morphological structure,
  such as the non-lexical or gesture-language
  fusion of classifier constructions, do not seem
  to be found in spoken languages
• Second, there is a core of morphological
  properties and processes that are found in all
  (?) sign languages