Please enable JavaScript.
Coggle requires JavaScript to display documents.
Categorical perception - Y2 - Coggle Diagram
Categorical perception - Y2
Recap of concepts and categories
Concepts are internal psychological representations that provide a means of understanding the world
Concepts can be defined as the building blocks of semantic knowledge
A way of categorising items, ideas and abstractions
Categories - a class of concepts that share some common properties
Our conceptual knowledge can be organised by means of categorisation
When we categorise, we group similar things together
Catergorisation is a fundamental feature of human cognition
Our senses are constantly bombarded by a wide array of information or concepts
One way of making sense of this information is to organise it into different categories
Fuzzy boundaries - Members of categories are not always entirely clear cut, with many items often having a boundary membership
McCloskey and Glucksberg (1978) suggested categories often form a continuum, with some items very clearly considered a member, some very clearly not, and some 'intermediate items'
-> 100% of people agree a dog is an animal, 0% believe a car is not and 47% believe yeast is not an animal
Typicality - some seen to be better or more clear cut than others: - Typical members have properties that are common to many category members
A penguin is less typical of a bird category than a robin
A shirt is more typical of clothing than a cape
Influence of typicality -
-> Typical members are rated as part of a category more often than atypical (Hampton, 1979)
-> Reaction times are faster to catergorise typical members compared to atypical members (Rips et al, 1973)
-> Age of acquisition is lower for typical members (Mervis and Pani, 1980)
-> Typical members are comprehended during language processing more quickly (Garrod and Sanford, 1977)
-> Typical members are usually first in a sentence before atypical members (Onishi et al, 2008 - people mention apples and lemons)
Categorical hierarchy -
Superordinate - e.g. animal, mammal, fish - all encompassing concepts with large membership
Basic - a typical member of that category which may describe a group e.g. dog
Subordinate - a specific category for one type e.g. a breed of dog
Categorical perception
Defined as an abrupt perceptual change at the boundary (Harnad, 2005) which can be seen in situations where the perceived change does not occur gradually but as instances of different categories
Categorical perception of colour is particularly good way of testing this, as there are variations in the way the world's languages segment colour space
Investigating colour categorical perception can help us understand where perceptual categories come from
-> Categorical perception (CP) of colour is shown when colours from the same category are discriminated less easily, quickly or accurately than equivalently spaced colours that cross a category boundary (Harnad, 1987)
-> Categories cannot be simply separated by environmental input - have to use some experience of shades for example to discriminate
Relativism v universality
Where do categories and their groupings come from - linguistic relativism (Whorf Hypothesis):
Categories are driven by language / based on linguistic structures
Language is a defining framework for categorisation, not just a way to describe them (Whorf, 1956) - speaking a language is agreeing to its categorisation of the environment
Perceptual categories are constructed through languages
Linguistic categories -> perceptual categories
Language, thought and colour - Reiger and Kay, 2009:
-> The Whorf hypothesis (linguistic relativity) that our language filters our semantic categories is right in two ways
Language influences colour perception primarily in half of the visual field
Colour naming across languages is shaped by both universal and language specific forces
Evidence -
Mandarin speakers conceive time vertically, whereas English speakers conceive it horizontally (Fuhrmann et al, 2011)
The Hanuxoo people in the Philippines have 92 different names for varieties of rice
There are hundreds of camel related words in Arabic
Inuit (Eskimo) people have 4 different words for snow
The Dani people of Indonesian New-Guinea have only two basic colour terms - mola for bright, warm colours and mili for dark, cold colours
Cross cultural influences in categorisation -
Colour naming systems (Reiger et al, 2015)
Robertson, Davidoff and Davies, 2000 - comparing colour perception in speakers whose language segment colour space differently - found Berinmo (Papua New Guinea) had no colour perception of blue and green, whereas English speakers do
Winawer et al, 2007 - compared English and Russian speakers
-> Russian shows CP on a forced choice categorisation task by English speakers did not
-> This categorical perception was eliminated by verbal interference, indicating language influences on CP
Lillo et al, 2018 - compared three dialect languages of Spanish (Castillian, Mexican and Uruguayan)
-> Found all 3 dialects had the same 11 categories
-> This was with the exception of Uruguayan speakers, who appeared to have a separate perceptual category for very light blue (celeste)
-> Although this suggests some innate categorisation, it also suggests more so that dialects with the same basic language categorise the same
Universality -
Are categories driven by perception - innate and hardwired
Berlin, 1991 - humans recognise the inherent order and structure of the biological world
Perceptual categories are hardwired and universal
Perceptual categories -> linguistic categories
Kay (2015) - Universality of Colour categories:
Different languages colour lexicons are more similar than they would be if each language partitioned the colour space idiosyncratically and that within these limits there is further order in the way languages acquire new colour terms
For example, some languages do not separate blue and green, and some do not separate yellow and red, but no language fails to do both
Suggests that language categorisation to inform top down perception of colour perception is innate in that all languages create colour categories, but the construction of the categories themselves are cultural
Summary -
Conceptual knowledge is represented in the form of categories
Categorisation helps us to make sense of the world by giving order to the array of information our senses receive
We can use the domain of colour to evaluate where perceptual categories come from
-> There has been much debate regarding whether perceptual categories are influenced by innate imperatives and by our learnt linguistic structures
-> Cross cultural studies suggest language affects CP - developmental research also suggests that innate structure can affect CP, but language can have an affect
Hemispheric evidence suggests we develop a LH bias in CP during colour term acquisition
Weak linguistic relativity - human factors like language and culture influence CP
Strong form - human factors like language and culture wholly determine CP
We have the innate ability to categorise, and language defines these
Language and concepts
Jackendoff (1987) - final general framework for integrating the linguistic and non-linguistic forms of knowledge
F-MIND - functional mind; a concept that suggests accounts are are occurring at a functional level
Division between external and internal worlds and within the internal world, the F-MIND contains various knowledge and processing systems
Noises <-> language <-> CS <-> SpS <-> haptic + visual + action system <- objects (objects go through sensory systems, which then encode the meaning of them and produce language to identify them)
Meaning has a CS and SpS component - CS is conceptual structure and SpS is spatial structure
SpS - spatial understanding of the physical world by integrating shape, motion and layout - where and when
-> Crucial in identifying real-world concrete objects - links to theory of representation and recognition of objects (Marr, 1982)
-> Distinguishes similar objects using shape and even understand hidden objects
CS - encodes category membership (ovoid cells in hippocampus - identifying familiar and unfamiliar objects) - what and how
-> Abstract objects of thought exist in the absence of SpS
Influences of categorisation on perceptual discrimination (Goldstone, 1994):
Stimulus generalisation - observer ability to make similar stimuli the same; this is how categories are formed
-> Type-token distinction - magazines are a type, the Sun would be a token; generalisation occurs when different tokens of the same type are seen as the same
-> Assigning tokens to relevant type based on relevant mental categories
-> Stimulus discrimination therefore involves the assigning to different categories
Categorical perception - easier and quicker to distinguish stimuli from different categories than stimuli from the same category (Liberman et al, 1957)
-> Suggestion that mental categories influence the manner in which we perceive the world
Goldstone (1994) - experiment on how people categories objects - Acquired distinctiveness - category judgements based on a relevant dimension become particularly high tuned
-> Subtle differences become easily detected when differences on the relevant dimension, such as size, when dealing with adjacent categories
-> Notman, Sowden and Ozgen (2005) - categorical perception expansion effect - space between adjacent categories expanded through training
-> Perceptual tuning - experience sharpens perceptual boundaries between stimuli taken from different categories
Acquired equivalence - training introduces blurring of boundaries
-> Becoming less adept at a irrelevant dimension, such as colour
-> However, there was no evidence for this in Goldstone’s study
Categorical perception and verbal labelling -
Do categorical perception effects reveal more about verbal labelling than perception - Goldstone, Lippa and Shiffrin (2001) suggested that adopting the strategy of labelling same or different leads to giving the same label to similar objects and a different label to others across a category boundary
Effect reveals description of the world rather than perception of it
-> Goldstone et al (2001) - facial perception
-> Trained participants to assign faces to either category A or B
-> Prior to training, similarity ratings were collected for the faces, rating how similar each face was to every other face in their set of four, and relative to a neutral face
-> Training given on assigning each face to a category, given distinctive feedback
-> Similarity judgement conducted again
Categorical perception expansion effect noted - faces assigned to different categories rated as less similar to one another following category training
However, variability compared to the neutral face reduced after training - different members of the same category had acquired equivalence
Category learning enhances representation of attributes that are critical to category membership - special status is given to attributes that are maintained well in the mind
Representational changes are different to language, and perception is not simply description
Roschian hypothesis - language does not determine perception, but is simply a system used to describe it
Developmental approach to categorical perception
Still an argument as to whether CP is dependent on colour language or whether CP can exist in the absence of colour language - in lieu with hemispheric approach
This can be addressed using the developmental approach
By using pre-linguistic infants, we can eliminate the problem that participants could be using labelling strategies when making colour category distinctions
Franklin and Davies (2004) - Do pre-linguistic infants (4 months) have colour categories -
Familiarisation - infants spend less time looking at a stimulus that is repeatedly presented
Paired comparison - novelty preference method
-> When presented with new and old stimulus, infants prefer novel stimuli
-> Baby in car seat, with two projectors projecting onto a stimulus window - video camera set up 40cm from baby recording them, projectors 107 cm from screen, light behind the camera in box
Habituation - phase 1 - time:successive trials showing them green
Test colours - phase 2 - test trials: identical; cross-category, within-category - green, blue, light green
Target colours were counter-balanced across trials
-> For example, one set of infants were habituated to Blue 2 then shown Blue 2 paired with Green 1 - between
-> Another set of infants habituated to Blue 2 then shown Blue 2 and Blue 1 - within
-> Looking times for Green 1 and Blue 2 compared
Familirisation results - mean stimulus times (in seconds) over ten familiarisation trials
Shorter fixations as habituated to stimuli in both colour sets (Red and pink, blue and purple)
Found that novelty preference was higher when the colours were within over between category, in all three colour trials
Cross cultural and developmental approach - Franklin et al, 2005:
Looking at toddlers wjho speak English and toddlers who speak Himba at the stage of colour term acquisition
-> Allows to test for effects of categorical perception developmentally, but also allows for a cross cultural comparison
-> Two alternative forced task modified for use with young children
-> Shown a bear wearing a coloured jumper, then covered it up, then revealed another bear wearing another jumper or the same (cross or within category foil)
-> Replicated the evidence of Franklin and Davies (2004)
-> Name, no boundary and reverse name boundary also tested - found linguistic knowledge did not impact CP
-> Both cultures displayed evidence of CP
Saji, Imai and Asano (2020) - Higher average score across cultures with age - all improved colour perception at similar times, with some colours that were subcategories of others being slower
Forbes and Plunkett (2020) -
-> Speakers of different languages develop the ability to use language terms at different rates
-> Suggests that while we do have innate categories, the use of these language differs by the nature of the language, such as frequency
Hemispheric approach
Colour categorical perception has also been tested by exploring hemispheric asymmetries in the effect - bias which changes with age
This approach has helped to clarify the role of language in colour categorisation
-> Brain is organised co-laterally
-> Visual projections from RVF processed in LH, where language function is lateralised
-> Visual projections from LVF processed in RH
-> If language is involved in colour CP, one might expect CP to be greater for stimuli in the RVF
Gilbert, Reiger, Kay and Ivry (2006) -
Colour CP - without interference, reaction times to within and cross category stimuli was similar in the LVF (right hemisphere), but within category took significantly longer in RVF than between category perception
-> With verbal interference, within category stimuli recognised slower in LVF (RH) than between categories, with the opposite (significant) effect in the RVF (LH)
-> Interestingly, the CP of cross categories was slower when interfered with in the RVF, with performance on within categories items being similar
-> Suggests language interference impacts the ability to decipher between category items, and thus language builds boundaries of categories
Animal recognition - LVF showed faster reaction to between categories, with a more enhanced effect in the RVF
Franklin et al, 2008 - used eye movements to compare adults and infants
Can use time to fixate a target (the different colour circle) as an indices of CP
Showed a left hemisphere bas for colour CP in adults and a right hemisphere bias of colour CP in infants
Right hemisphere bias for colour CP in toddlers still learning colour terms, and left hemisphere bias for colour CP in toddlers who have learned colour terms - suggests perception interacts with language, and until colour language is learned, to categorise, visual stimuli is relied upon
Left hemisphere allows perception using language terms - language defines categories and describes categories learned in visual stages
Maier and Abdel Rahman, 2019 -
Fixation cross (1000ms) -> visual search array (200 ms) -> fixation cross until reaction (no longer than 2300ms)