Sign Language Morphology II

1
Sign Language Morphology II

• COGS 524, Cognitive and Linguistic aspects of
  Sign Language
• 21 March 2007
• Sandler Lillo-Martin 2006, chapters 5,6,7

2
Classifiers in spoken language

• Example from Caddo, a language of North America
  (Glück and Pfau, 1997)
• Kapi kan-cân'i-ah
• Coffee Clfluid-buy-PERF
• 'He bought (fluid) coffee'
• Kapi dân-cân'i-ah
• Coffee Clpowder-buy-PERF
• 'He bought (ground) coffee'
• With the two classifiers kan- and dân- two
  classes of objects that can be bought are picked
  out fluid and ground objects.

3
Warming up An Example of a classifier
construction in Sign Language

• 1. Human-dragging-a-4-legged-creature-behind-him

• What properties does this CL contruction have ?
• Unit (N, compound, phrase)?
• Which phonological parameters are involved?
• On which level is it located? (phonological,
  morphological, syntax, discourse)

4
CL construction, continued

• What properties does this CL contruction have ?
• Unit? Word? Phrase? CL construction
• Which phonological parameters are involved?
  Handshape, orientation, movement
• On which level is it located? morphological,
  syntax, discourse

• What semantics do those features convey?
• Unit proposition
• Phonological parameters
• Handshape 1 two-legged being, 1 four-legged
  being
• orientation upright, walking orientation,
• movement moving forwards, dragging behind
• On which level is it located story-telling,
  narrative

5
Variability of classifier (CL) constructions

• All sign languages have a rich set of classifiers
  and make use of it pervasively
• Classifiers pick out classes of objects to which
  they refer collectively
• There are various kinds of classifiers bound,
  free, nominal, adjectival, verbal, subject,
  object, lexicalized, frozen
• They are realized as classifier handshapes,
  movements, orientations, locations.
• They can be used creatively in discourse and
  heavily draw on iconicity.

6
3 groups of classifiers

• 1. Entity/Class/semantic classifiers verbal
  classifiers that modify the verb in terms of its
  intransitive subject.
• 2.Handle classifiers verbal classifiers that
  modify the verb in terms of its transitive
  objects
• 3. SASS (Size-and-shape specifiers) nominals
  (like entity CL) or adjectival modifiers of a
  noun that visually depict the geometric shape of
  the object
• In all 3 types of CL the hand is depictive of
  physical or more abstract semantic properties of
  objects (Aronoff, Meir, Padden, and Sandler 2003)

7
1. Class/Entity/semantic classifiers

• Entity classifiers are used with subject nouns of
  intransitive, non-agentive verbs, e.g.
• Subjects of intransitive-agentive verbs
• CAT WALK-CL4-legged
• PERSON WALK-CL2-legged
• Subjects of intransitive-non-agentive verbs
• BALL BOUNCED-CLbig
• PENCIL ROLL-CLsmall

(Zwitserlood, 2003, 146)
8
Entity classifiers (from Sign Language of the
Netherlands, NGT)The G-hand CL

• G-Hand CL for noun referents that denote long,
  thin objects (Zwitserlood 2003 91)

9
Entity CL The flathand CL

• The flathand is used for referents whose shape is
  perceived as flat and wide (Zwitserlood 2003 93)

10
Flathand CL

• The Flathand may also serve as a classifier for
  the hands and the feet.

Zwitserlood 2003, 164
11
Manner of motion CLwith the flathand

• The elephant flies by moving its ears up and
  down (Zwitserlood 2003, 169)

12
Entity CL The claw-hand CL

• The claw-hand CL is used for spherical entities
  (Zwitserlood 2003, 97)

13
Entity CL The F-hand CL

• The F-hand CL is used for small, round, and thin
  referents such such buttons or coins
  (Zwitserlood, 2003, 99)

14
Generalized, default entity CL

• Zwitserlood argues that there is a default CL in
  NGT, the G-hand, that can refer to any kind of
  entity, even if the referents have quite
  different shapes, like a bird, an egg, or a fox
  (2003, 110).

15
Combinations of entity CLS

• A cup (CL-sperical) is standing next to a
  newspaper (CL-flat,wide)
• ISL, Aronoff, Meir, and Sandler, 2005, the
  Paradox of Sign Language

16
There are many more Entity-ClassifiersYou may
look at video-clips of them at the below website

• Entity classifiers are characterized by
  particular handshapes and hand orientations

http//www.jal.cc.il.us/ipp/Classifiers/
17
2. Handle CLs

• The very same classifying handshapes used as
  entity-CL may be used as handle-CLs, also. In a
  handle-CL, the CL changes in dependence of the
  direct object of the transitive verb.

18
Handle CLs the flathand CL

• The flathand-CL is used with verbs that have
  large and bulky direct objects (Zwitserlood 2003,
  95)

19
Handle CLs the claw-hand CL

• The claw-hand CL is used with verbs that have
  spherical direct objects (Zwitserlood 2003, 97)

20
Handle CL the F-hand CL

• The F-hand CL is used for verbs that have round,
  small, thin objects (Zwitserlood, 2003, 100)

21
PICK-UP and various handle CLs
http//www.ling.upenn.edu/courses/Summer_2004/ling
001/images/asl_pickup3.jpg
22
Generalized, default handle CL

• As with entity Cls, there seems to a exist two
  default handle CLs in NGT.
• However, there seem to be individual preferences
  which one to choose.

23
What's the differenceEntity vs. Handling CL

• Whether a CL is an entity CL (left-hand side) or
  a handling CL (right-hand side) depends on the
  argument structure of the verb.

24
Entity vs. Handling CL

• Entity classifiers (direct representation of
  entities) occur on intrransitive verbs of motion
  and location of a referent. Handling classifiers,
  on the other hand, represent manipulation of the
  referent in question and occur only on transitive
  verbs. Thus, the choice between an entity
  classifier and a handling classifier is
  determined by the argument structure of the
  verb. (Zwitserlood, 2003, p 127)

Zwitserlood's dissertation can be downloaded
chapter-wise from http//igitur-archive.library.
uu.nl/dissertations/2003-0717-122837/inhoud.htm
The quotation above is from http//igitur-archiv
e.library.uu.nl/dissertations/2003-0717-122837/c4.
pdf
25
The function of CL

• It has been argued that CL function as
• Lexical function CL as stems CL can act as
  stems of signs that are entered in the mental
  lexicon, e.g. WRITE, PERSON,
• Grammatical function CL as AGR markers Entity
  and handling CL agree with previously introduced
  subjects or objects

26
CL as AGR
Classifiers as agreement markers Two types of
classifiers, viz. entity classifiers and handling
classifiers, are systematically linked to
arguments of verbs of motion and location
(Shepard-Kegl 1986, Benedicto Brentari to
appear, Glück Pfau 1998, Zwitserlood 2003).
After introduction of a referent, it is possible
to leave that referent unexpressed in the
subsequent discourse, and its reference can be
backtracked by means of a classifier on the verb,
in particular verbs that express the motion of a
referent through space, its localization in
signing space and its existence in space.
(Zwitserlood, 2004).
27
Body vs. Hand CLs(Aronoff et al. 2003)

• ASL has many hand CL, i.e., the handshape and
  orientation convey the classifier construction.
• ISL has more referent projections The head and
  the whole body are used as referent projections
  in polymorphemic classifer construction,
  e.g.DOG PAWS-UNDER-CHIN
• The dog is sitting
• CAT LYING-ON-SIDE 2-LEGS-EXTENDED
• The cat is lying on the ground

28
ASL vs. ISL classifiers
ASL Hand classifier
ISL Referent projections
ISL Referent projections

• (Aronoff et al 2003)

29
Manner of motion CLS

• Zwitserlood 2003,

30
3. SASS CLs

• In an SASS, the handshape represents
  visual-geometric properties of the referent. An
  SASS CL can be refered to later by a handle CL
  later, e.g.
• PAPERSASS I-GIVE-YOUFLAT
• The size and shape introduced by SASSs can act as
  referent points for later verb classification.

31
2 kinds of SASS'S static and tracing(Supalla
1982)

• 1. Static SASS's show the outlines of a class of
  objects with one or both hands. They are the same
  as entity/object/class/semantic classifiers!

(Zwitserlood 2003, 151)
32
2 kinds of SASS'S static and tracing(Supalla
1982)

• 2. Tracing CLS are also called extent or
  perimeter CLs.
• In tracing SASS's the hands have a specific
  handshape and trace the outlines of a class of
  objects (with one or both hands). The movement
  brings about the classifier, not just the
  handshape, as in static SASSs.

(Zwitserlood 2003, 151)
33
Localization with static and tracing SASSs

• Through a classificatory movement, a tracing CL
  is placed at a particular locus in signing space

• An entity CL is placed at a particular locus in
  signing space

Zwitserlood 2003, 154
34
Elicitation of tracing SASSs in NGT

• Zwitserlood 2003, 159 ff

35
How signers render the figure in NGT1st variant
36
How signers render the figure in NGT2nd variant
37
CL constructions Verbs of motion, location, and
existence

• 1. verbs of motion show the path movement of a
  referent through signing space
• 2. verbs of location locate a referent in signing
  space
• 3. verbs of existence express the existence of a
  referent in space.
• (Zwitserlood 2003, 143ff)

38
1. verbs of motion
This motion verb has (one or two morphemes
referring to the start and the end of the
path and one for the entity (human).
This CL verb has one morpheme for
the orientation change and one for the entity
(human).

• (Zwitserlood 2003, 146f)

39
Verbs of motion in discourse

• The referents in CL constructions need not always
  be overt. They can also be dropped and act as
  implicit discourse antecedent to which the
  following CL verbs refer

Zwitserlood, 2003, 148)
40
2. verbs of location
This verb of location has one morpheme for the
locus and one for the entity (vehicle)

• Zwitserlood, 2003, 148

41
3. verbs of existence

• Verbs of existence are somewhat indistinguishable
  from verbs of location because as soon as a
  location is specified there is more to them as
  the mere existence.
• Thus, Zwitserlood gives the bicycle example as an
  example for a verb of existence, which is not
  quite justified.
• In DGS, there is a general existence predicate
• SCH which, however, is not a classifier. In
  TID, there is VAR.

42
The puzzle of CLs

• SL-M (2006) note that CL constructions are
  highly anomalous
• These forms are anomalous at every level of
  analysis. They are iconic yet conventionalized,
  at once mimetic and linguistic. (Sandler and
  Lillo-Martin 2006 17).
• In what respects are they anomalous?

• They are highly iconic. Therefore, they may not
  belong to the grammar of Sign Language at all but
  more to the gesture system
• They violate phonological constraints such as the
  symmetry and the dominance constraint . They are
  post-lexical and therefore do not preserve
  structure.
• They can freeze and become regular signs in the
  lexicon but melt again and become re-analyzed
  as classifiers again.
• They carry propositional meaning

43
Theoretical approaches to CLs

• Historically, there have been various approaches
  to CL and CL constructions/predicates
• 1. Classifiers are morphemes, CL predicates are
  to be analyzed morpho-syntactically. They are to
  be studied with regular linguistic tools.
• 2. Classifiers belong more to the gestural system
  and should not be counted as proper morphemes.
  They are outside language and linguistics.
• 3. Hybrid approaches CLS are somewhere between
  language and gesture

44
1. CLs as morphemes

• CLs are proper morpho-syntactic phenomena, that
  is they can be accommodated to spoken language
  linguistics.
• Hence, Sign Language is like Spoken Language.
• Each classifier handshape, movement, and
  orientation is a separable, digital morpheme, CLs
  can be decomposed like any other polymorphemic
  sign. Their overall meaning is compositional.
• CL show agreement, which is a proper linguistic
  function.
• Supalla 1982, Aronoff et al. 2003, 2005, Glück
  and Pfau 1997

45
2. CLS are gestural

• Since CLS are iconic and spatial phenomena, they
  are not to be analyzed linguistically, but as
  gestures or visual templates
• CLS are analogue, continuous, gradient, not
  digital, arbitrary, categorical, there is a lot
  of variation, improvization
• If we count every CL-information as a proper
  morpheme, there will be an inflation of
  morphemes.
• Hence, Sign language is not like Spoken Language,
  CLS are special to Sign Languages
• Liddell, Cogill-Koez, Schembri

46
3. Hybrid accounts

• CL constructions are controversial, unresolved,
  puzzling
• They have properties of and behave like both
  language and gesture at different times.
• Every language has such face-to-face
  representational resources (Johnston 2006), only
  spoken languages underexploit them
• We have to distinguish between language in a
  narrow and wide sense
• Johnston, Liddell 2003, Herzig and Emmorey

47
A representative of a hybrid account (with
great sympathy for the gestural part)

• Liddell's (2003) evaluation of Supalla and his
  own approach
• Supallas proposal was based on the idea that all
  meaning must come from morphemes. I suggest an
  approach in which some meaning comes from
  identifyable morphemes, some meaning is
  associated with the full lexical unit itself, and
  meaning is also constructed by means of mental
  space mappings motivated by the variable and
  gradient ways that the hand is located and
  oriented. (Liddell 2003 273-4)

48
CL as templated visual representation(Cogill-Koez
2000)

• She claims that in sign language, two fully equal
  channels of representation of propositional
  information are used, the linguistic and the
  visual. Classifier constructions (CPs) are not
  linguistic, but highly abstract, schematic visual
  representations built up from discrete parts
  which she calls templates, in a combinatorial
  fashion.

49
What are visual templates?

• A template is a form that occurs repeatedly, over
  various contexts, with the same physical
  realization. Some of these templates are discrete
  but not digital. They allow for elastic
  analogue depictions of handshapes, orientations,
  movements and locations. Representing Classifier
  Predicates by TVRs avoids the trouble a fully
  linguistic analysis encounters. It integrates the
  iconic, analogue, and gestural mode, yet it
  maintains the idea of a combinatorical,
  compositional construal of meaning. By claiming
  that TVRs are not language, but intriguingly
  language-like, Cogill-Koez underlines their
  liminal (borderline) character.

50
The dynamic nature of CLS between 'freezing' and
'melting'

• CLS may commute between the Lexicon and the TVR
  system. Both systems are equally important in
  SLs.
• (Cogill-Koez 2000)

51
Frozen CLs The lexicalization of CLS

• CLs may become fully lexicalized, i.e., the
  handshape has lost its morphological character
  and is not flexible anymore but has merely
  phonological quality. (Aronoff et al. 2003, 69)

52
Frozen CLs The lexicalization of CLS

• The frozen CL sign BOOK has the same form in
  many SLs. The previous CL handshapes depicting
  the pages of the book have lost their
  morphological status.
• (Sandler and L-Martin 2006, in Keith Brown)

53
Watching the freezing/lexicalization of CLs
on-line
Aronoff et al. (2003, 73f)

• At first, the signer produces an extended
  description using Cls for showing the ligament.
• During some trials, he comes up with a much more
  concise form using a single CL construction.

54
The melting of frozen forms

• ISL sign WRITE, has once been a classifier
  construction.
• It can, if discourse favours it, become
  re-analysed again as such. Then, e.g. The signer
  may pantomimic-like look at what has been
  written on the page.
• Crossing the borders between the lexicon and CL
  predicates happens easily and frequently.

(Aronoff, Meir, Padden, and Sandler 2003)
55
The proportion of discrete linguistic structure
and gestural potential in the ontogeny of Signed
and spoken languages

• SL equal proportion of linguistic and gestural
  part
• SpL Linguistic part stronger than gestural part
• (Cogill-Koez 2000)

56
Empirical studies on Classifiers

• Emmorey Herzig (2003) studied whether location
  and handshape classifying morphemes are processed
  in a digital, categorical or in a gradient,
  continuous way.
• Deaf and hearing subjects were tested.

57
Exp. 1 Interpretation of Gradient Variation in
Location
58
Subjects had to place a sticker in relation to a
bar as indicated by the signer (left-hand
pictures).In this CL construction, the
position of the dot with respect to the reference
line was varied in 30 pictures
59
Result classified location

• Deaf and hearing subjects do not differ in their
  dot placement. (Both have a tendency to move away
  from the central vertical axis)
• --gt There is gradient, non-linguistic processing
  of location

60
Experiment 2 Interpretation of gradient
variation in handshape to indicate size

• In this eperiment, subjects hat to rate the size
  of a medallion

61
Gradient variants of ASL CL- handshapes (F, baby
C) as input for a judgement task on the actual
size of a medallion
62
10 continuous response possibilities for judging
the medallion size as shown by the ASL signer
63
Results

• Hearing subjects responded incoherently

• Deaf subjects responded with continuously growing
  sticker size to the CL handshapes they had seen
  (in random order)--gt linguistic effect of
  categorically perceived CL handshapes. They know
  that an F-handshape indicates a very small
  round object whereas a C handshape indicates a
  large one. They also know which aspect of the CL
  morphemes convey gradient, analogue information
  (small vs. normal F small vs. wide baby C)

64
Conclusions from Emmorey Herzig (2003)

• Loc is not morphemic
• Locations in signing space are mapped in an
  analogue way to the physical location of objects
  in space by both deaf and hearing subjects. There
  are no placement morphemes as Supalla had
  suggested.
• Handshape is morphemic
• Size as conveyed by continuously varying
  handshape CLs is judged through the knowledge of
  CL handshape by deaf subjects but not by hearing
  subjects. Deaf subjects perceive the CL
  handshapes categorically (small, medium, big) but
  also acknowledge gradient information.

65
References

• Sandler, Wendy and Lillo-Martin, Diane (2006)
  Sign Language and linguistic universals.
  Cambridge CUP
• Aronoff, Mark, Meir, Irit, Padden, Carol, and
  Sandler, Wendy. (2003). Classifier Complexes
  and Morphology in Two Sign Languages. (PDF) In
  Karen Emmorey (ed.), Classifiers in Spoken and
  Signed Languages. Mark Aronoff, Irit Meir, and
  Carol Padden, and Wendy Sandler. 53-84.
• Glück, S. R. Pfau (1998) On Classifying
  Classification as a Class of Inflection in German
  Sign Language. In Cambier-Langeveld, T., A.
  Lipták M. Redford (Eds.), Proceedings of
  Console VI. pp. 59-74.
• Zwitserlood, Inge (2004) The functions and use of
  sign language classifiers. In TISLR 8 Barcelona,
  September 30 - October 2. Programme and
  Abstracts. (International Conference on
  Theoretical Issues in Sign Language Research 8)
  (2004)

66
References

• Cogill-Koez, Dorothea (2000a). Signed language
  classifier predicates Linguistic structures or
  schematic visual representation? Sign Language
  and Linguistics, 3 (2), 153-207.
• (2000b). A model of signed language classifier
  predicates as templated visual representation.
  Sign Language and Linguistics 3 (2), 209-236.
• Emmorey, Karen and Herzig, Melissa (2003).
  Categorical versus gradient properties of
  classifier constructions in ASL. In Perspectives
  on classifier constsructions in sign languages,
  Karen Emmorey (ed.), 222-246. Mahwah, NJ
  Lawrence Erlbaum Associates.