Decomposing Posture Verbs into Frame Attributes Thomas Gamerschlag - - PowerPoint PPT Presentation

decomposing posture verbs into frame attributes
SMART_READER_LITE
LIVE PREVIEW

Decomposing Posture Verbs into Frame Attributes Thomas Gamerschlag - - PowerPoint PPT Presentation

Decomposing Posture Verbs into Frame Attributes Thomas Gamerschlag & Wiebke Petersen Institute of Language and Information Heinrich Heine Universitt Dsseldorf 4th UK Cognitive Linguistics Conference, London, 10 July 2012 Posture verbs


slide-1
SLIDE 1

Decomposing Posture Verbs into Frame Attributes

Thomas Gamerschlag & Wiebke Petersen Institute of Language and Information Heinrich Heine Universität Düsseldorf 4th UK Cognitive Linguistics Conference, London, 10 July 2012

slide-2
SLIDE 2
  • Posture verbs (e.g. stand, sit, kneel) encode

gestalt properties and spatial information, which are perceived by gestalt recognition and spatial orientation.

→ Posture verbs (PVs) are an excellent object of

the investigation of cognition and language (like spatial prepositions and dimensional adjectives). Posture verbs (PVs)

slide-3
SLIDE 3

The garden gnome is sitting on the mushroom. figure posture locational ground relation

General scenario

slide-4
SLIDE 4
  • The meaning and use of PVs and other positional verbs is well

described for German: (Berthele 2004, Gerling & Orthen 1979, Kaufmann 1994, 95; Kutscher & Schultze-Berndt 2007, Lang & Carstensen 1990, Maienborn 1990, 91; Serra Borneto 1996, Schönefeld 2006, Wunderlich & Kaufmann 1990 among others)

PVs in German

slide-5
SLIDE 5

stehen 'stand' liegen 'lie'

Orientation matters

slide-6
SLIDE 6

Der Gartenzwerg steht auf dem Sockel. 'The garden gnome is standing on the plinth.' Der Gartenzwerg liegt auf der Wiese. 'The garden gnome is lying on the lawn.'

Orientation matters

slide-7
SLIDE 7

support from below

Support matters, too!

stehen 'stand'

slide-8
SLIDE 8

support from above

Support matters, too!

hängen 'hang'

slide-9
SLIDE 9

Der Gartenzwerg hängt am Seil/am Galgen. 'The garden gnome is hanging from the rope/the gallows.'

Support matters, too!

slide-10
SLIDE 10

sitzen 'sit'

The supported part (of the body) also matters

bottom knees

knien 'kneel'

slide-11
SLIDE 11

Der Gartenzwerg sitzt auf der Schaukel. 'The garden gnome is sitting on the swing.' Der Gartenzwerg kniet auf der Wiese. 'The garden gnome is kneeling on the lawn.'

The supported part (of the body) also matters

slide-12
SLIDE 12

'Standing' and 'lying' do not necessarily involve specific parts of the body

slide-13
SLIDE 13

...or other objects

slide-14
SLIDE 14

Support matters, too!

nor does 'hanging'

slide-15
SLIDE 15

'Standing' and 'lying' are dependent on inherent gestalt properties

#stehender / liegender Ziegelstein #standing / lying brick stehender / (?) liegender Sarg standing / (?) lying coffin

canonical vertical

slide-16
SLIDE 16

'Standing' and 'lying' are dependent on inherent object properties

canonical vertical maximal extent vertical

stehender / # liegender Sarg standing / # lying coffin

slide-17
SLIDE 17

Object schema for brick (Lang 2001): 1D 2D 3D Max Across Min 1D 2D 3D Max Across Min (canonic.) Vert Spatial knowledge of objects can be captured in object schemata (Lang 1987, 2001) Object schema for coffin (Lang 2001): salience/prominence of axes

slide-18
SLIDE 18

Representation of PVs in two-level semantics (Kaufmann 1995)

  • a. knien 'kneel': λPλx [KNEEL(x) & P(x)]
  • b. Int(KNEEL(x)) = ∃y [supportf (d-us(y), knee(x))]
  • a. stehen 'stand': λPλx [STAND(x) & P(x)]
  • b. Int(STAND(x)) = ∃y [supportf (d-us(y), s(prom(x)))]
  • a. liegen 'lie': λPλx [LIE(x) & P(x)]
  • b. Int(LIE(x)) = ∃y[supportf (d-us(y)), s(nprom(x)))]
  • a. hängen 'hang': λPλx [HANG(x) & P(x)]
  • b. Int(HANG(X)) = ∃y [supportf (d-nus(y), side(x))]
slide-19
SLIDE 19

knien 'kneel': Int(KNEEL(x)) = ∃y [supportf(d-us(y), knee(x))]

knees underside

  • f knees

upper side of supporting ground knees

Representation of knien 'kneel' in two-level semantics (Kaufmann 1995)

slide-20
SLIDE 20

20

Analysis: Framework

  • Conceptual knowledge is captured in frame

representations.

  • Frame representations provide an explicit,

variable-free, and cognitively plausible format

  • Frames are defined as recursive attribute-

value structures (Barsalou 1992).

  • The attributes correspond to mathematical

functions.

slide-21
SLIDE 21

Der Gartenzwerg kniet. 'The garden gnome is kneeling.'

  • a. knien 'kneel': λPλx [KNEEL(x) & P(x)]
  • b. Int(KNEEL(x)) = ∃y [supportf(d-us(y), knee(x))]

theme physical state gestalt garden gnome knees

under- side supporter

solid

  • deict. upper

side kneeling gestalt

slide-22
SLIDE 22

Der Gartenzwerg ist/befindet sich auf der Wiese. 'The garden gnome is/is located on the lawn.'

sein / sich befinden 'be / be located'

upper region

slide-23
SLIDE 23

upper region garden gnome lawn upper region knees

under- side supporter

solid

  • deict. upper

side kneeling gestalt physical state

Der Gartenzwerg kniet auf der Wiese. 'The garden gnome is kneeling on the lawn.'

  • a. knien 'kneel': λPλx [KNEEL(x) & P(x)]
  • b. Int(KNEEL(x)) = ∃y [supportf(d-us(y), knee(x))]
  • c. auf 'on':[-DIR]: λyλx [LOC(x, UPPER_REGION(y)) & CONTACT(x,y)]
slide-24
SLIDE 24

upper region garden gnome lawn upper region knees

under- side supporter

kneeling gestalt

Der Gartenzwerg kniet auf der Wiese. 'The garden gnome is kneeling on the lawn.'

  • a. knien 'kneel': λPλx [KNEEL(x) & P(x)]
  • b. Int(KNEEL(x)) = ∃y [supportf(d-us(y), knee(x))]
  • c. auf 'on':[-DIR]: λyλx [LOC(x, UPPER_REGION(y)) & CONTACT(x,y)]

solid

slide-25
SLIDE 25

Der Gartenzwerg sitzt auf der Schaukel. 'The garden gnome is sitting on the swing.'

upper region garden gnome swing upper region bottom

under- side supporter

sitting gestalt solid

  • deict. upper

side physical state

slide-26
SLIDE 26

Object schema of Ziegelstein 'brick'

gestalt 1D 2D 3D maximal axis across axis minimal axis salience/prominence of axes: 1D > 2D > 3D brick 1D 2D 3D Max Across Min

slide-27
SLIDE 27

Der Ziegelstein liegt auf dem Tisch. 'The brick is lying on the table.'

liegen 'lie': maximal axis (1D) = horizontal lying bricks

slide-28
SLIDE 28

Der Ziegelstein liegt auf dem Tisch. 'The brick is lying on the table.'

lying gestalt upper region garden gnome table upper region supporter solid physical state

  • deict. upper

side deictic underside horiz.

  • rientat.

1D

slide-29
SLIDE 29

Der Ziegelstein steht auf dem Tisch. 'The brick is standing on the table.'

standing bricks stehen 'stand': minimal axis ≠ vertical

slide-30
SLIDE 30

standing gestalt upper region garden gnome swing upper region supporter solid physical state

  • deict. upper

side deictic underside

stehen 'stand': option 1

horiz.

  • rientat.

3D

  • ption 1:

minimal axis (3D) ≠ vertical → minimal axis = vertical

slide-31
SLIDE 31

Der Ziegelstein steht auf dem Tisch. 'The brick is standing on the table.'

upper region brick table upper region supporter solid physical state

  • deict. upper

side deictic underside standing gestalt horiz.

  • rientat.

3D

slide-32
SLIDE 32

gestalt 1D 2D 3D maximal axis across axis minimal axis saliency/prominency of axis: 1D > 2D > 3D coffin

Object schema of Sarg 'coffin'

canonical vertical

slide-33
SLIDE 33

Der Sarg steht in der Kapelle. lit.: 'The coffin is standing in the chapel.'

canonical vertical = vertical

slide-34
SLIDE 34

upper region garden gnome swing upper region supporter solid physical state

  • deict. upper

side

stehen 'stand': option 2

deictic underside standing gestalt vert.

  • rientat.

canonic. vertical

  • ption 1:

canonical vertical = vertical

slide-35
SLIDE 35

Der Sarg steht in der Kapelle. 'The coffin is standing in the chapel.'

interior region coffin chapel interior region supporter

canonic. vertical

floor

3D

deictic underside standing gestalt horiz.

  • rientat.

solid physical state

  • deict. upper

side

slide-36
SLIDE 36
  • All the elements necessary for the analysis of

PVs (support, object axes, …) are captured by frame representations → no extra-representational format

  • The combinatorics of a PV and the elements

co-occurring with it is captured technically by the unification of the frames of the parts. Advantages of a frame analysis of PVs

slide-37
SLIDE 37
  • The representation format allows for easy

access to the relevant object properties

  • Frame representations show a flexible degree
  • f complexity (zooming in and out by

expanding/not expanding nodes) Advantages of a frame analysis of PVs