Embodiment Flashcards
How do we mentally represent the meanings of words?
It’s not known, but the two main views are the traditional view and the embodied view.
Outline the traditional view of representation.
Meanings are represented as features, usually described in terms of other words, like a dictionary.
What is a problem with the traditional view of representation?
The symbol grounding problem - meanings need to be grounded in the physical, social and emotional world, as demonstrated by Harnad (1990)’s Chinese dictionary example. Much of the learning of words (abstract symbols) consists of relating them to objects and actions which have manifestations in the real world and are accessible through perception and action.
Outline the embodied view of representation.
We learn word meanings by associating them with real objects and actions through memory - when an object and a word co-occur, they are associated with memory traces over time resulting in memory simulation and the creation of perceptual representations.
According to the embodied view, what perceptual features are represented?
Visual orientation, object shape and visibility.
According to the traditional and embodied views, how do people represent “the pencil is in the cup?”
Traditional view = amodal symbols, manipulated using rules, e.g. [IN[PENCIL,CUP]]
Embodied view = people should mentally represent the orientation of the pencil based on their experiences. Therefore there should be information on orientation.
What did Stanfield and Zwaan (2001) do?
A sentence-picture verification task to test whether visual orientation is represented. Had participants read the sentence “John put the pencil in the cup” or “John put the pencil in the drawer”, then showed them a picture of a vertical or horizontal pencil and had them decide whether the item was mentioned in the sentence.
What did Stanfield and Zwaan (2001) find?
Reaction times were faster when the object orientation was consistent with mental representations from the sentence. This supports the embodied view and mental representations of orientation.
What did Zwaan, Stanfield & Yaxley (2002) do?
Investigated mental representation of object shape using the same sentence-picture verification task and design as Stanfield and Zwaan (2001), whereby participants see an egg in a shell or a cracked egg (no visible shell) and then must decide whether the item was mentioned in the sentence “Mary saw the egg in the eggbox” and “Mary saw the egg in the frying pan”.
What did Zwaan, Stanfield & Yaxley (2002) find?
Reaction times were faster in consistent conditions, suggesting that we mentally represent object shape.
What did Yaxley & Zwaan (2007) do?
Investigated mental representation of visibility using the usual sentence-verification task design in which participants saw a high or low resolution image of a moose and had to decide whether the item was mentioned in the sentence “through clean goggles, the skier could easily identify the moose” or “through the fogged goggles, the skier could hardly identify the moose”.
What did Yaxley & Zwaan (2007) find, and what does this suggest?
Faster reaction times in consistent conditions, supporting the idea that visibility (contextual influence) is also represented. This suggests a detailed mental representation and a strong interaction between perception and language.
What did Wassenberg & Zwaan (2010) do?
Investigated whether pictures affect sentence comprehension - the opposite way round than the sentence-picture verification tasks. Three phases (presented as unrelated experiments):
Phase 1: word-picture verification task - participants shown the word Toothbrush and a picture of one (half participants shown horizontal, half vertical), and asked whether the picture matches the word.
Phase 2: Mental rotation filler task
Phase 3: eye-tracking experiment using the sentence “Aunt Karin finally found the toothbrush in the sink/cup beside the mirror”.
What did Wassenberg & Zwaan (2010) find?
First pass reading times were greater for when the sentence in phase 3 didn’t match the toothbrush orientation in phase 1 but only at the point of the sentence where the mismatch was i.e. ‘in the cup’ or ‘in the sink’. Consistent = shorter first pass reading times.
What do Wassenberg & Zwaan (2010)’s findings suggest?
That pictures affect mental representations and therefore sentence comprehension.
What did Coppens, Gootjes & Zwaan (2001) do?
Three studies (presented as unrelated experiments):
Phase 1: word-picture verification task using an unfolded or folded ironing board.
Phase 2: emotional word stroop task (15mins)
Phase 3 (ERP): sentences presented word-by-word while EEG signals were recorded using the sentence “Esther was looking for kitchen steps. She pulled open the hall closet. In the closet was an ironing board”.
What did Coppens, Gootjes & Zwaan (2001) find?
In the mismatch condition ERPs showed N400.
What did Hauk, Johnsrude & Pullvermuller (2004) do?
Using fMRI, studied brain activation for words - first had participant make movements (foot, finger and tongue) in order to measure their typical responses. Wanted to see whether action related words may also activate the same motor areas in the brain, so had participants read action words (e.g. kick) in isolation which aimed to activate similar areas in the brain (leg, arm and face words).
What did Hauk, Johnsrude & Pullvermuller (2004) find?
The words did activate similar motor areas in the brain. This supports embodiment, as it suggests that words are linked to mental representations of actions.
However there was less of an effect for face words, perhaps because tongue movements were used as the control.
What did Raposo et al. (2009) do?
Investigated action words in context to see whether the word alone is linked to motor area activation. Used idioms and literal sentences to compare against hand/foot movement activation.
What did Raposo et al. (2009) find?
No activation in motor areas for idiomatic sentences, therefore activation depends on context, rather than just the word alone.
What is a problem with Raposo et al. (2009)’s findings?
fMRI has an approximately 5 second delay, and might not have been sensitive enough to pick up activation for idiomatic sentences.
What did Glenberg & Kaschak (2002) do?
Investigated the action-sentence compatibility effect by asking participants to decide whether sentences involving movements towards or away from oneself made sense by pressing buttons which were either further or closer than the middle button, which they held while reading the sentence.
What did Glenberg & Kaschak (2002) find?
Participants are faster to move their hands in the direction that’s implied in the sentence.
