Size Matters Sara C. Sereno Patrick J. O’Donnell Margaret E. Sereno.

Size Matters

Sara C. Sereno

Patrick J. O’Donnell

Margaret E. Sereno

Bigger is better

• Large vs. small visual object– Activation of more neurons– Attract attention more easily– May hold attention for longer

Bigger is better

• Ethology– Mate selection (e.g., alpha males)

– Supernormal stimulus

(Tinbergen & Perdeck, 1950)

Bigger is better

• Size-value effect (Bruner & Goodman, 1947)

50 20>

Bigger is better

• Size-congruity effects– Pavio (1975)

– Rubinsten & Henik (2002)

+ +

+ ZEBRALAMP + ZEBRALAMP

Bigger is better

• Line bisection with numbers (Fischer, 2001)

91

28

Bigger is better

• Linguistic markedness– Unmarked = usual, dominant, basic, default form– Marked = (not the above)– Examples:

Gender marking: general/male female

lion, actor lioness, actress

Size: How tall is X?

How big is y?

How wide is z?

Semantic Size

• Hypothesis– Words denoting “big” entities are easier to process

than those denoting “small” entities.– RTs for semantically “big” words will be faster than

those for semantically “small” words.

Lexical Decision Experiment

• Subjects: N=28– 14 female, 14 male– 26 years old– right-handed

• Apparatus– Mac G4 using PsyScope 1.2.5 PPC software– 24-pt Courier font (black on white)– 3 characters = 1o vis. angle

Lexical Decision Experiment

• Materials– Big/Small defined in relation to human size

N Length Syl FreqImageability

Big 45 6.20 2.00 24.40 6.08

Small 45 6.20 2.00 23.74 6.07

Frequency BNC (occurrences per million)

Imageability MRC Psycholinguistic Database

(scale 1-7) Bristol Norms (Stadthagen-Gonzalez & Davis, 2006)

Cortese & Fugett’s (2004) Imageability Norms

– 90 length-matched pseudowords (e.g., blimble)

Materials

BIG SMALL BIG SMALL BIG SMALLbed cup truck thumb buffalo apricotbay fly whale peach gorilla parsleyjet lip camel glove giraffe emeraldcow pin comet snail mountain magazinepark rose moose tulip motorway bacteriatree neck planet button elephant moleculebear ring jungle needle wardrobe sandwichlake nose galaxy insect dinosaur parasitetank tape rocket bullet downtown mosquitobull leaf walrus peanut bookcase teaspoonriver glass monster diamond cathedral cigarettetrain phone stadium battery submarine butterflyhorse video mansion vitamin skyscraper fingernailocean apple tractor sausage supermarket handkerchiefshore robin volcano aspirin hippopotamus hummingbird

Lexical Decision Experiment

• Procedure– Instructed that words represent a selection of several

different categories of objects.

NW W– Response mapping:

left right

1000 ms 500 ms200 ms

+

until response

string

Results

• Data exclusion– Overall: RTs < 250 ms & RTs > 1500 ms– Per subj per cond: RTs < –2SD & RTs > +2SD– 4.72% data loss

Results

RT (ms) %Err

Big words 513 (8.6) 1.6 (.3)

Small words 527 (9.3) 2.3 (.5)

t1(27)=5.22, p<.001 ts<1.15, ps>.25

t2(44)=3.29, p<.01

Discussion

• Potential confound of response mapping:– Spatial markedness

• Right is for WORD response, Big or Small.

• Spatially, however, Left is marked and Right is unmarked.

• Consistency of markedness (Right, Big) confers benefit only to Big words.

– SNARC (Daheane et al., 1993)• Spatial Numerical Association of Response Codes

• Faster right-sided responses to larger numbers; faster left-sided responses to smaller numbers.

• E.g., Which is bigger? vs.4 · 9 9 · 4

888888888888888888888888888888

22222222222222222222222

Discussion

• Test potential confound: W NW

– Reverse response mapping:

leftright

– Subjects: N=14 (7F,7M), 23 years old– Materials: identical– Procedure: identical– Results: 5% data exclusion

Discussion

• Replication: RT (ms) %ErrBig words 514 2.4

Small words 527 3.1 t1(13)=2.71, p<.05 ts<1.05,

ps>.30

t2(44)=2.08, p<.05

• Combined data: RT (ms)Big words 513

Small words 527 F1(1,40)=28.12, p<.001

F2(1,88)=12.72, p<.001

Discussion

• Is size coded in lexical representations?– Yes, for size words and for some like dwarf, giant.

• Is size a feature of concrete nouns?– Yes, according to size-congruity studies.

• However, these studies use a size comparison task.

– Yes, according to current Lexical Decision results.• But, response criteria can still play a role.

Possible Explanation

– Larger objects contain more Low SF information.– Low SF is transmitted faster thru magno pathway.– 1o vis cortex & LGN are activated during imagery.– If imagery accompanies word recognition, this

information may become available earlier for words referring to larger objects.

– Thus, while both Big and Small items can be equally highly imageable, the relative speed of accessing a stored visual representation is faster when the object is bigger.

Conclusion

• Need to establish effect in other paradigms:– EM-reading in neutral context.– EM-reading in different contexts (e.g., large ant).– EEG, MEG, fMRI, & WHATNOT.

• The Bottom Line…………………..

Bigger is FASTERFASTER