proportion correct answers

Pro

porti

on c

orre

ct a

nsw

ers

Subject IDS1 S2 S3 S4 S5 S6 S7 S8 S9 S10

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Proportion correct answers across subjects

"To bite, or not to bite“: Perceived causality in the perception of negative partsPatrick Spröte & Roland W. Fleming | Justus-Liebig Universität Gießen

BackgroundOur data suggest that subjects are good at inferring the causal history of unfamiliar ‘bitten’ 2D shapes.On a between subject basis the relative depth of a negative part & and its relative area are good predictor of the subjects’ judgements.

Subjects can do the task Subjects are confident Differences in judgements

The more concave, the more ‘bitten‘

The larger the missing portion, the more ‘bitten‘

The more angles, the more ‘bitten‘

60° 90° 108° 135° 152° -5.31 -3.67 -2.51 -0.95 0.72

Decomposition Negative Parts

Stimuli were created from convex, irregular hexagons. From half of the stimuli, a portion of the shape was deleted by random intersection with another hexagon and removing the region of overlap.

Whole

Bitten

Stimulus statisticsα β

Relative area Mean of interior angles

Width of aperture

Relative depth

We presented different shapes in different sizes and orientations. Subjects indicated with a cursor on a 10-point scale the extent to which each object appeared to them to be ‘bitten‘.

Time

Interior angles are no longerpredictive

Relation to figure/ground

The investigated shape properties in Experiment 1 were highly inter-correlated. Therefore we conducted a second Experiment with a set of stimuli for which, by design, the correlation between “mean of interior angles” & “relative depth” was kept near zero.

Relatability

patrick.sproete@googlemail.com

Stimuli: “Hexagon world”

Task

Results (Experiment 1)

Results (Experiment 2)

Discussion

Genericity

Decreasing Symmetry

counting vertices

concave = bitten

mean

Rel

ativ

e fre

quen

cy

Ratingbittenwhole

Relative frequency categories were used

1 2 3 4 5 6 7 8 9 100

0.02

0.04

0.06

0.08

0.1

0.12

0.14

0.16

Pro

porti

on c

orre

ct a

nsw

ers

(who

le s

hape

s)Proportion correct answers

(bitten shapes)

Interindividual differences injudging ‘bitten‘ vs. ‘whole‘ shapes

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Mean of the interior angles55 65 75 85 95 105 115 125 135 145 155

1

2

3

4

5

6

7

8

9

10

Rat

ing

bitte

nw

hole

Average rating as a function of the mean of the interior angles

log relative depth-6 -5 -4 -3 -2 -1 0 1

1

2

3

4

5

6

7

8

9

10

Rat

ing

bitte

nw

hole

Average rating as a function of the logarithm of relative

depth

-6 -5 -4 -3 -2 -1 01

2

3

4

5

6

7

8

9

10

log relative area

Rat

ing

bitte

nw

hole

Average rating as a function of the logarithm of relative

areaR²=.63

R²=.76

R²=.90

55 65 75 85 95 105 115 125 135 145 1551

2

3

4

5

6

7

8

9

10

Mean of the interior angles

Rat

ing

bitte

nw

hole

Average rating as a function of the mean of the interior angles

-6 -5 -4 -3 -2 -1 01

2

3

4

5

6

7

8

9

10

Average rating as a function of the logarithm of relative depth

log relative depth

Rat

ing

bitte

nw

hole

Negative parts‘ concavity predicts ‘bites‘

R²=.10

R²=.59

-9 -8 -7 -6 -5 -4 -3 -2 -1 01

2

3

4

5

6

7

8

9

10

Average rating as a function of the logarithm of relative area

R²=.33

log relative area

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1R² (mean interior angles) as a

function of R² (log relative depth)

R² (log relative depth)

R² (

mea

n of

the

inte

rior a

ngle

s)

Mean of interior angles loses predictive power when

uncorrelated with relative depth

90° 108° 120° 135° 140°129° -5.22 -3.88 -0.42-1.77-2.6 -8.76 -5.75 -3.4 -1.2 -0.64

-5.6 -3.74 -2.73 -1.64 -0.54