Michael Leo Glynn
Does a thought need language for it to exist?
Research in the area of the development of cognition and language has a long tradition; yet, until recently it has not been possible to ascertain how infants think as they are not able to communicate verbally. It could be argued that language provides concepts that are used to organise thinking, and this premise suggests that infants are neither able to think, nor possess knowledge before they learn language. However, recent research employing innovative experimental methods, allows inferences into how infants’ minds function before they acquire language. This essay will outline relevant research and highlight some methods used to examine how preverbal infants think and categorise the world before they can talk, and will challenge the view that infants have not developed the ability to conceptualise before language acquisition. It will also examine how both language comprehension and language production develops, and discuss how infants are able to distinguish speech. In doing so, it will acknowledge nativist and empiricist perspectives, whilst considering conflicting views of developmental theorists.
Perspectives conflict in the way in which it is assumed cognition develops on a continuum from innate ability to experiential learning. Theories offered by Chomsky (1965) in relation to language-specific mechanisms, for example, are from a nativist viewpoint. In contrast, behaviourists, e.g., Skinner (1954) take an empiricist stance advocating that development is primarily a result of learning (Oates and Grayson, 2004). Piaget (1955) and Vygotsky (1962) hold similar views towards the constructive nature of cognitive development and the role that language plays, agreeing that language is elaborated through complex interactions between an infant and their environment. However, Piaget submits that language is dependent upon fundamental constructs of thought for development, whereas, Vygotsky takes a social constructivist view and sees language as necessary for developing thought; furthermore, that thought and language are two separate functions that merge at around 2-years old (Bancroft and Flynn, 2005). Piaget observed infants talking to themselves, which he termed ‘egocentric speak’ (Oates and Grayson 2004), a symbolic function that enables infants to internally construct verbal thought. Vygotsky, in contrast, viewed this as only the first step, and that social interaction with others using language as a ‘cultural tool’ is crucial for language development (Oates and Grayson, 2004).
Research has continued apace since Piaget and Vygotsky’s studies and, whilst their influences remain, there is now compelling evidence that preverbal infants can establish links between experiences, construct categories, group them, and in doing so form concepts much earlier than previously assumed. Based upon Frantz’s (1963) familiarisation/novelty preference method, whereby visual fixation duration for one stimulus over another confirmed that infants can perceptually categorise, Younger and Gotlieb (1998) conducted controlled experiments with infants aged 3, 5, and 7 months. Infants were familiarised with distorted prototype dot patterns ranging from good, intermediate, to poor before being shown a control pair comprising a previously shown non-distorted exemplar and an unfamiliar novel exemplar. With one exception all groups preferred to focus upon the novel prototype, thereby indicating that they had formed a category representation due to the familiarisation of the distorted set of exemplars. Results suggested that infants are able to organise their thinking, and although the greatest ability to preferentiate was present at 7 months, all infants possessed a degree of cognition.
Quinn et al. (1993), using pictures rather than patterns, arguably providing more ecological validity, demonstrated that infants are also able to categorise animals: discriminating cats from other species such as birds, dogs and horses and, furthermore, able to discern related species (Eimas and Quinn, 1994). Behl-Chadha’s (1996) experiments revealed that infants aged 3 to 4 months are also capable of forming hierarchical structures similar to adults. A novelty preference method using familiar furniture objects as a ‘class’, at a global level, revealed that infants are able to distinguish between chairs and couches at a basic/intermediate level. Experiments using photographs of mammals found that infants, in common with adults, are also able to form global category representations for wide-ranging classes of stimuli (Quinn and Oates, 2004). Further studies by Younger and Gotlieb (1998) indicate that infants use two processes to store categories, an ‘exemplar memory’ for small numbers of instances where every example is stored, and a ‘prototype abstraction’, for large numbers of instances whereby an average of the examples are stored. Their findings suggest that infants employ two strategies to categorise across global/superordinate, basic/intermediate, and specific/subordinate levels.
Object examination experiments sought to identify whether infants construct or deconstruct from basic to global levels of category representations or vice versa, and findings tend to support that the latter occurs easier and earlier (Quinn and Oates, 2004), suggesting that prototype abstraction develops before exemplar memory. Studies by Quinn (1994), using similar methods illustrate that infants are also capable of spatial category representation, e.g., above as opposed to below, and between rather than outside. Sensor modality cues are also an important factor in how infants form category representations and there has been much interest in how infants focus upon specific attributes that provide these cues. Quinn and Eimas (1996b), amongst others, employed methods that systematically varied the attributes of an exemplar shown to infants which revealed that they formed categories based upon the presence of a cue, but were unable to form category representations in its absence (Quinn and Oates, 2004). Rakison and Butterworth (1998) employing a sequential touching procedure, the categorisation of toy objects, with older infants, also reported evidence that specific aspects of an exemplar are a cue that enables differentiation between global categories. Using dynamic point light display methods, Arterberry and Bornstein (2002) found that dynamic movement cues can also be attributed to how infants form category representations, where infants were able to distinguish, by movement, animals from objects.
Two competing theories exist as to how categories mature into concepts. In accord with the Piagetian viewpoint, a single-process model suggests that language, amongst other exemplars of information, contribute towards an infant’s cognitive ability to develop category representations: a process described as ‘quantitative enrichment’ (Quinn and Eimas, 2000), e.g., enrichment of category representations by infants learning to name objects with their caregivers (Quinn and Oates, 2004). Mandler (1997) addressed the issue of whether there is a developmental progression from category to concept and theorised that perceptual categorisation, the ‘knowing’ that something exists, occurs before conceptual categorisation, the forming of concepts including ‘thinking’ and ‘understanding’ (Quinn and Oates, 2004). Mandler (1992, 2000) suggests a two-process model arguing that perceptual and image schemas, in parallel, lay the foundation for mature concepts. Perceptual schemas describe features, whereas image schemas allow for abstract features of how exemplars behave leading to the formation of true concepts. This accords with Paiget’s theory of structures and stages of development being marked by characteristic modes of thought (Bancroft and Flynn, 2005). A similar dual-process theory, Karmiloff-Smith (1986), describes an infant’s developmental process as knowledge moving from being implicit and procedural to explicit. Karmiloff-Smiths ‘representational re-description’ model suggests that practise at procedural level is required for knowledge to be re-described as an ‘object of thought’.
Research supports that preverbal infants have learned how to form categories, possess a capacity for complex recognition strategies, and have the basic cognitive building blocks for language to develop. However, it is the point at which language begins to emerge, the nature of the relationship between category learning and identifying word sounds in speech, understanding them, categorising them, and reproducing them that is debatable. Speech production is dependent upon comprehension which requires: identification of a word from a speech stream; remembering the word sound for recall; association with an object or action; repetition; then using the word in an appropriate context (Harris, 2004). This empirical viewpoint suggests language needs to be learnt and is not innate. However, DeCasper and Spence (1996) found that prenatal infants are capable of recognising speech sounds; furthermore, research indicates that 4-week-old infants show a propensity towards their mother’s voice (Mehler and Dupoux, 1994). Experiments by Mehler et al. (1994) report that younger infants favour the familiar language of their surroundings. Christophe and Morton (1998) language comparison experiments attributed this phenomenon to a preference for prosodic patterns, predicting and finding that 2-month-old infants could distinguish English from Japanese, but given the similarities between English and Dutch could not do so. The ability of infants to exploit prosodic cues to bound words was further investigated by Johnson and Jusczuk (2001), who elaborated that transitional probabilities, the ability to detect and remember syllables, also contributes toward language comprehension.
Infants are capable of category learning prior to language acquisition, therefore it seems plausible that category learning must be achieved before language production, and it is at this point that changes in an infant’s developing vocal track support the production of their first words, whilst at the same time learning to categorise words and relate them to objects, e.g., the formulation of a word category for cat and then labelling the category. Vygotsky’s social constructivist viewpoint holds that relating words to objects and actions occurs within a social context. Bruner (1975, 1993) elaborates the view that first words emerge within familiar social contexts; however, he did not proffer the extent to which they grow out of experience. The empirical view supports the notion that the production of words involves repeating the sound of a word as it is experienced and in an appropriate context. Harris et al. (1983) support Vygotsky’s view of embedding language in socially meaningful activities, and found that infants use words in a similar way to their mothers during social routines; furthermore, that comprehension is enhanced by qualitative enrichment: clarifying actions such as gazing, pointing and touching objects, e.g., a toy cat, as cues for an infant to relate words to objects (Harris, 2004). However, as comprehension of words improves, infants are able to abstract them in other contexts. Hart (1991) attributed this non-reliance on experience to increased vocabulary ability.
A vocabulary spurt occurs between 1 to 2 years and appears crucial to understanding the correlation between language development and categorisation. Gopnik and Meltzoff (1987) sought to substantiate a link between categorisation and language domains of development and whether they occurred simultaneously. Longitudinal studies observing infants between the ages of 15 to 20 months reported a significant relationship between the vocabulary spurt and categorisation, but did not find strong links between categorisation and other cognitive measures. Gopnik and Meltzoff (1992) were unable to establish whether the vocabulary spurt causes an improved ability to categorise or whether, indeed, the ability to categorise underpins language development (Harris, 2004). These findings compliment Vygotsky’s theory that language and thought are separate functions, and supports that until the point at which they merge infants only possess a basic comprehension of language which provides the building blocks for higher-level thinking and the realisation of ‘true concepts’ (The Open University, 2006, p. 11).
Evidence has identified that infants possess a degree of cognitive ability that allows for a basic understanding of language, albeit a rudimentary recognition of basic prosodic patterns and syllables. Prior to language acquisition infants are capable of forming, storing, and retrieving category representations and recognising cues, at both basic and global levels using exemplar memory and prototype abstraction. Category learning can also be distinguished from other cognitive skills as beneficial to language development and when infants start to use their first words, it is apparent that their ability to categorise visual objects evolves to forming abstract ideas. Moreover, as vocabulary increases the capacity to categorise provides for the transition from knowing to forming true concepts. Prior to contemporary methods of investigation, nativists may not have been able to attribute these capacities to anything but innate ability and recent approaches challenge this view. The capacity to categorise is not dependent upon language acquisition; rather it facilitates the foundation for language development. Notwithstanding that infants possess the physical and cognitive processes requisite for developing language, also essential are social interactions that provide rich social contexts in which they can develop cognition and language. This supports Vygotsky’s premise of language being a cultural tool that features significantly in the development of cognition.
Arterberry, M. E. and Bornstein, M. H. (2002), cited in Quinn and Oates (2004) p. 42.
Bancroft, D. and Flynn, E. (2005) ‘Early cognitive development’, in Oates, J., Wood, C. and Grayson, A. (eds), Psychological Development and Early Childhood, Oxford, Blackwell/The Open University.
Behl-Chadha, G. (1996), cited in Oates and Grayson (2004) p. 36.
Bruner, J. S. (1975), cited in Harris (2004) p. 73.
Bruner, J. S. (1993), cited in Harris (2004) p. 73.
Christophe, A. and Morton, J. (1998), cited in Harris (2004) p. 67.
Christophe, A. and Morton, J. (1998) ‘Reading A: Is Dutch native English? Linguistic analysis by 2-month-olds’ in Oates and Grayson (2004) pp. 98-105.
Chomsky, N. (1965), cited in Oates and Grayson (2004) p. 14.
DeCasper, A. J. and Spence, M. J. (1996), cited in Harris (2004) p. 66.
Eimas, P. D. and Quinn, P. C. (1994), cited in Quinn and Oates (2004) pp. 34-35.
Fantz, R. (1963), cited in Quinn and Oates (2004) p. 28.
Fantz, R. (1963) ‘Reading A: ‘Pattern vision in newborn infants’ in Slater and Oates, (2005) pp. 124-6.
Gopnik, A. and Meltzoff, A. N. (1987), cited in Harris (2004) p. 54-55.
Gopnik, A. and Meltzoff, A. N. (1992), cited in Harris (2004) p. 56.
Harris, M., Jones, D. and Grant, J. (1983), cited in Harris (2004) p. 74-76.
Harris, M. (2004) ‘First words’, in Oates, J. and Grayson, A. (eds), Cognitive and Language Development in Children, Oxford, Blackwell/the Open University.
Hart, B. (1991), cited in Harris (2004) p. 89.
Johnson, E. K. and Jusczyk, P. W. (2001), cited in Harris (2004) p. 70-72.
Mandler, J. M. (1997), cited in Quinn and Oates (2004) p. 27.
Mandler, J. M. (1992, 2000), cited in Quinn and Oates (2004) p. 44.
Mehler, J. and Dupoux, E. (1994), cited in Harris (2004) p. 67.
Mehler, J., Jusczyk, P. W., Dehaene-Lambertz, G., Dupoux, E. and Nazzi, T. (1994), cited in Harris (2004) p. 67.
Oates, J., Sheehy, K. and Wood, C. (2005) ‘Theories of development’ in Oates, J., Wood, C. and Grayson, A. (eds), Psychological Development and Early Childhood, Oxford, Blackwell/The Open University.
Oates, J. and Grayson, A. (2004) ‘Introduction: perspectives on cognitive and language development, in Oates, J. and Grayson, A. (eds), Cognitive and Language Development in Children, Oxford, Blackwell/the Open University.
Piaget, J. (1923/1926), cited in Oates, J., Sheehy, K. and Wood, C. (2005) p. 66.
Piaget, J. (1955), cited in Oates and Grayson (2004) p. 17.
Quinn, P. C. (1994), cited in Quinn and Oates (2004) p. 37.
Quinn, P.C. and Eimas, P. D. (2000), cited in Quinn and Oates (2004) p. 43.
Quinn, P.C. and Eimas, P. D. (2004b), cited in Quinn and Oates (2004) p. 40.
Quinn, P. C. Eimas, P. D. and Rosenkrantz, S. L. (1993), cited in Quinn and Oates (2004) pp. 34-35.
Quinn, P. C. and Oates, J. (2004) ‘Early category representations and concepts’ in Oates, J. and Grayson, A. (eds), Cognitive and Language Development in Children, Oxford, Blackwell/the Open University.
Rakison, D. and Butterworth, G. (1998), cited in Quinn and Oates (2004) p. 41.
Skinner, B. F. (1953), cited in Oates and Grayson (2004) p. 14.
The Open University, (2006), ED209 Child Development, Study Guide, Milton Keynes, The Open University
Vygotsky, L. S. (1962), cited in Oates and Grayson (2004) p. 17.
Vygotsky, L. S. (1986), cited in Bancroft and Flynn (2005) p. 72.
Younger, B. A. and Gotlieb, S. (1988), cited in Quinn and Oates (2004) pp. 31-33.