Discuss the method of ‘familiarisation/novelty preference’ and consider its contribution to psychologists’ understanding of infant cognition.
One of the keystones in an infant’s development is the ability to group similar items and experiences together. This, on the surface, may seem a trivial skill but it forms the basis for much of the infant’s cognitive development in the first months of life. Once similar things are identified into groups, structure and order can form around them. This process is referred to as ‘categorisation’. In fact, the development of the process itself provides a useful insight into the developmental progression of an infant more generally. One of the major tools psychologists have used to study this phenomenon is that of the ‘familiarisation/novelty preference’ technique:
Fantz (1963) noted that infants showed a strong tendency to pay attention to novel objects, compared with those they had previously encountered. If an infant is presented with an object for an extended period of time, that infant will gradually reduce the attention it bestows upon the object. The baby will begin to look away, until eventually it no longer pays the object attention at all. This process is known as familiarisation (or ‘habituation’). Subsequently, if the infant is presented with the same object as before, alongside a new object (with which the infant has had no previous experience), then vastly more attention will be paid to the novel item. This is called ‘novelty preference’. Presumably this pairing of phenomena (familiarisation and novelty preference) comes about from a biological tendency (genetically hard-wired) which ensures an infant experiences as much of its environment as possible, in order to learn at an optimum rate.
Psychologists have produced a lab-based version of the ‘familiarisation/novelty preference’ phenomenon in order to examine its effect on the cognitive development of infants. The technique has two stages: In stage 1 an infant is shown a number of different objects belonging to the same category (e.g. Siamese cat, Persian cat, Tabby cat). In the second stage the infant is presented with a pair of novel stimuli. One of the stimuli belongs to the category the baby has just encountered (e.g. Manx cat), the other belongs to an entirely new category (e.g. Labrador dog). The infant is then graded on the preference they pay to each stimulus. The infant normally shows a greater preference for the stimulus from the novel category. This is because they have formed a representation of the familiar exposed category (i.e. cats) which became habituated, so more examples of this category will hold less attention. When a new category is encountered (i.e. dogs) more attention is due to this novel item since it has not been seen before. This process is used to examine many aspects of infant development which relate to categorisation, for example: how do infants form categories? How are these categories remembered? How are they organised? Also, since categorisation and language formation are so strongly linked the ‘familiarisation/novelty preference’ technique is used to understand the development of language in infants.
In order to examine this process simply some of the first studies of infant cognition resorted to using very basic stimuli. Younger and Gotlieb (1988, see also Bomba and Siqueland, 1983; Quinn, 1987) used simple dot patterns, known to be effective in examining adult categorisation. Infants were familiarised with six pairs of distorted dot patterns, which had been derived from a single (undistorted) prototype (and hence were considered to belong to the same category). The infants were then shown a test pair which included the prototype of the exposed category, and a prototype of an entirely different dot category. The amount of time the infants spent looking at the novel dot pattern was recorded. When the prototypes were very simple patterns the infants (aged 3-7 months) spent a significantly larger amount of time observing the novel prototype than the familiar one. This indicated that they had all formed a representation of the dot category, without ever seeing the pattern which defined the category (the prototype). As the prototype patterns became more complex only older infants (5 months and above) showed this significant trend. So, older infants appeared to be better at forming a prototype from the series of distorted examples, although all showed evidence of category formation.
Younger and Gotlieb (1988) went on to use this finding to examine how infants actually store their category representations. They hypothesised two possibilities for category storage: 1. all possible exemplars are stored in memory and are available for comparison with new instances (‘exemplar memory’); 2. an average of all observed exemplars is stored as a prototype (‘prototype memory’). Initially it would seem plausible that prototype memory is the more likely as this is the most efficient form of storage and retrieval. Comparing an example with all previous examples would be very time-consuming. Once another cohort of Younger and Gotlieb’s (1988) infants had been exposed to the distorted dot pattern exemplars (see above) they were then shown the prototype paired with one of the previously seen distorted exemplars. If the infant produced a prototype when it was exposed to the exemplars earlier (by averaging the features of the distorted patterns) then the prototype the infant formed should look much like the actual prototype. In this case the infant should perceive the distorted pattern as less familiar (and thus attended to for more time) than the prototype. If the infant was in fact just remembering each and every pattern it was presented with then the distorted exemplar should be more familiar (and attended to less) than the prototype, which was not seen until this point.
However, it seems that infants use both of these category storage mechanisms, depending on the exact parameters of the experiment (i.e. if there are a few simple exemplars then it is more efficient to encode each one, when there are many complex exemplars a prototype is more appropriate). More importantly, infants use the same mechanism as adults performing the equivalent test. This not only indicates that infants are able to form prototypes (an essential mechanism for category formation), but are capable of adult-like cognitive tasks from a very early age (ED209, Child Development Course Team, 2008).
Experiments like those described above have been criticised for their lack of environmental validity. In order to address whether or not infants can actually categorise items that are relevant to their surroundings a number of authors have used the ‘familiarisation/novelty preference’ technique. Quinn, Eimas and Rosenkrantz (1993, see also Eimas and Quinn, 1994; Quinn and Eimas, 1996) gave infants exposure to pictures of domestic cats from different breeds and in different orientations. Subsequently, the infants spent less time viewing novel cat pictures (as they considered them familiar) than pictures of animals from other species (which belonged to novel categories). These experiments show that infant categorisation is reproducible outside of the ‘lab’. Moreover, the experiments indicate that infants can produce categories that are both environmentally valid and useful, without the assistance of a vocabulary.
Knowing that similar things go together is the first stepping stone to producing useful categorical knowledge. The next step than an infant makes is to organise their categories into hierarchical structures. This step brings the child closer to forming a strong basis for a lexical framework (i.e. towards speech). To illustrate: A Siamese cat belongs to the super-ordinate category of ‘cats’, which in turn are ‘animals’. Construction of this categorical framework is commonly investigated using the ‘familiarisation/novelty preference’ technique:
Behl-Chadha (1996) set out to discover if infants were truly able to form hierarchical structure. Infants aged between 3 and 4 months were familiarised with a set of twelve photos of chairs (which included sub-ordinate categories like desk chairs and rocking chairs etc.). Following this the infants were shown pictures of novel chairs along with other items of furniture. This infant paid more attention to the novel items than the chair-related items. This standard ‘familiarisation/novelty preference’ effect showed that the babies had successfully formed the category ‘chair’. However, when the infants were familiarised with a set of ‘couch’ pictures they subsequently treated new ‘couch’ pictures as familiar, but pictures of other chair-types were treated as novel, attracting more attention (indicating the babies knew couches were an individual category, whilst at the same time knowing that chairs were a category also). This experiment proved that infants are in fact able to ‘nest’ categorical information into a hierarchical structure, needed for the formation of a vocabulary.
Another aspect of categorical grouping that is a pre-requisite of early speech formation is that of spatial relation. This form of categorisation is more abstract than the types summarised above as it cannot rely on perceptual features. Quinn (1994, see also Quinn et al., 2003) showed that infants can categorise abstract spatial relations, grouping objects that are ‘above’ or ‘below’. If an infant was familiarised with stimuli that were all of the same spatial relation they would subsequently show preference for stimuli in another spatial relation. These kinds of experiment show that infants produce seemingly complicated categorical information without the a priori powers of speech and vocabulary. Infants therefore have the cognitive ability to form many complex representations of their environment. In fact, many authors believe this forms the basis for communication and language development.
Waxman and Markow (1995) suggest that language acquisition is promoted due to the ability it provides the infant in referring to objects. In fact the onset of speech and the so-called ‘vocabulary spurt’ have both been attributed to categorisation. Gopnik and Meltzoff (1992), for example, note that children who are better at categorisation on the ‘familiarisation/novelty preference’ test are those who also use more words and names for items in their first months of speech production. Goldfield and Reznick (1990), note that half of all early words spoken by infants were object names, further strengthening the link between object category and cognitive development, and language in particular. The ‘familiarisation/novelty preference’ method is therefore key in understanding the building-blocks of infant cognition and speech.
Bomba, P. C. and Siqueland, E. R. (1983) ‘The nature and structure of infant form categories’, Journal of Experimental Child Psychology, vol. 35, pp. 294–328.
ED209 Course Team (2008). Cognitive and Language Development in Children, Milton Keynes: The Open University.
Eimas, P. D. and Quinn, P. C. (1994) ‘Studies on the formation of perceptually based basic-level categories in young infants’, Child Development, vol. 65, pp. 903–17.
Fantz, R. L. (1963) ‘Pattern vision in newborn infants’, Science, vol. 140, pp. 296–7.
Gopnik, A. and Meltzoff, A. N. (1992) ‘Categorization and naming: basic-level sorting in eighteen-month-olds and its relation to language’, Child Development, vol. 63, pp. 1091–103.
Quinn, P. C. (1987) ‘The categorical representation of visual pattern information by young infants’, Cognition, vol. 27, pp. 145–79.
Quinn, P. C. (1994) ‘The categorization of above and below spatial relations by young infants’, Child Development, vol. 65, pp. 58–69.
Quinn, P. C. and Eimas, P. D. (1996) ‘Perceptual organization and categorization in young infants’, Advances in Infancy Research, vol. 10, pp. 1–36.
Quinn, P. C., Eimas, P. D. and Rosenkrantz, S. L. (1993) ‘Evidence for representations of perceptually similar natural categories by 3-month-old and 4-month-old infants’, Perception, vol. 22, pp. 463–75.
Quinn, P. C., Adams, A., Kennedy, E. et al. (2003) ‘Development of an abstract category representation for the spatial relation ‘‘between’’ in 6-to 10-month-old infants’, Developmental Psychology, vol. 39, pp. 151–63.
Younger, B. A. and Gotlieb, S. (1988) ‘Development of categorization skills: changes in the nature or structure of infant form categories?’, Developmental Psychology, vol. 24, pp. 611–19.
Waxman, S. R. and Markow, D. B. (1995) ‘Words as invitations to form categories: evidence from 12-to 13-month-old infants’, Cognitive Psychology, vol. 29, pp. 257–302.