It might seem possible to predict a child’s first words to some extent, particularly for that child’s caregiver, as they are aware of the kinds of words spoken to their infant and which of those words the infant responds to the most. However, it is more of a challenge to predict the way in which the child will attempt to both produce, and understand the complexity of meaning of these words. Although the child may have some understanding of, and respond to, many words that they are exposed to, it is another task entirely for the child to produce the word in the correct semantic context. Infants also need to learn to distinguish between individual sounds and learn to produce them correctly in order that their speech is comprehendible. Linguists have studied early children’s speech and found particular patterns relating to aspects of lexical development. In this essay I will be discussing to what extent the aspects of semantics and phonology are predictable within a child’s early spoken language and what predictions could be made about these particular areas of lexical development.
The types of words first spoken in infancy are arguably predictable; not the particular words themselves but the category in which they fall under. A child’s first word tends to be first spoken between 10- 14 months of age (McLean & McLean, 1999 p.139) and their earliest words are ones referring to “people, objects, actions, attributes, states, and relationships” (McLean & McLean, 1999 p.139) that the child encounters within their personal environment. The environments in which children are brought up in tend to be generally similar, e.g. with a caregiver, in a home, with a bed, food, drink and toys. However, these environments are individual, specific to different families, therefore there will be much variation between them (McLean & McLean, 1999 p.139). McLean and McLean suggest that, although there are differences between infants’ environments, the nature of their first words is hardly different in that they show attention to and engage with “action events”; therefore the objects involved in the actions and the descriptions of the actions themselves tend to make up the early vocabulary of a child (McLean & McLean, 1999 p.140). Examples of these might be ‘ball’, ‘doll’, ‘cup’ and ‘push’, ‘up’, ‘down’. Bloom in her 1970 research stated that the word choices of children’s one-word utterances came under one of two categories: Substantive forms or Functional/Relational forms. According to Bloom, substantive forms are object and action labels, and functional/relational forms state or relationship descriptions relating to these objects or events (Bloom, 1970). Given this observation, it seems fair to argue that a child’s lexical development in terms of the category of word choice is predictable; however, predicting infants’ specific early vocabulary appears to be a fairly impossible task provided that each child has a unique learning environment. Therefore, the impossibility of appreciating/recognising the particular nature of every individual environment denies the ability to predict the first words an infant will speak.
Further research into early speech comes from Katherine Nelson in 1973 who looked at the early speech of 18 infants aged between 12-24 months and recorded the first 50 words they produced. She noted two most common forms produced by the infants: the ‘general nominal’ and the ‘specific nominal’ (Nelson, 1973). The ‘general nominal’ describes nouns- being the most common among the infants’ early speech- and the ‘specific nominal’ refers to proper nouns- the second most common form produced by the infants. Bloom had identified similar word choices to Nelson with nouns being most prominent, although, by analysing the speech of her daughter, Bloom found that function words featured more in her initial one-word utterances. A personal example of my cousin’s early speech shows support of this research as among her earliest one-word utterances was the functional/relational word (preposition) ‘down’, along with ‘more’ (pronoun). This difference in Nelson’s and Bloom’s results proposes further evidence for the impossibility of predicting the first words produced by infants; yet, it provides support for Mclean and McLean’s idea that children engage with and therefore produce vocabulary relating to actions and the objects involved.
Within the bracket of lexical development is the idea of the child understanding the meaning of the words they are exposed to and begin to produce themselves. Their home environments alone are packed with different and complex concepts that the child has to learn. The child may have learnt the word, but are they using it in the correct semantic sense? From the age of 8 months, infants begin to make the connections between the words they hear and their referents (P.J. Brooks & V. Kempe, 2012, p.70) By the age of 2 years making associations between words and their referents happens much more efficiently (Fernald, Perfors & Marchman, 2006). Since words are bound to specific contexts and a child might hear a word in the same situation repeatedly, the error of under-extension is often made, particularly nearer the beginning of their lexical development (Barret, Harris & Chasin, 1991). A child may learn a word and use it correctly in one context, yet they might not understand its relevance in another context and therefore not use it due to the lack of learnt association in this context. The child is yet to understand the complexity of the word’s meaning; treating it as a linear concept. An example of this might be an infant making the association between the word ‘milk’ and their bottle of milk they have at bedtime. They understand the meaning of ‘milk’ in this context, however the sense of ‘milk’ in a carton in the fridge has not yet been realised; they do not recognise anything other than their bedtime bottle as being associated with the word ‘milk’. This type of error, though, often diminishes as “cross-situational” learning occurs- the child takes notice of the word being used in multiple situations, then learns to use the word in this broader sense (Smith & Yu, 2008).
Opposite to that of under-extension errors, associating a word with objects that are not within the same specific category is called over-extension. One reason that has been proposed for a child making this type of error is that there are gaps in an infant’s vocabulary and so they use a word that they already know that they think relates in some sense to the object they wish to talk about (Becker, 1994b: Clarke, 1982, 1993: Swan, 2000). A personal example of this type of error comes from my cousin who learnt the word ‘apple’, yet associated it with similarly shaped fruit. For example, she would see a tangerine or a plum in a fruit bowl and say “apple”, over-extending the meaning of the word to cover all round fruit of similar size. Due to the amount of research into, and evidence of, under- and over- extension errors it seems fair to say that it could be predicted that infants will make these types of errors and even possibly at around what age. However, it is very unlikely that one would be able to predict which particular words a child may incorrectly label (over-extension) or not associate with the word at all (under- extension) as these errors depend on the child’s own reasoning for under- or over- extending a word. What one child may consider semantically similar, another child may not. For example, my cousin over-extended the word ‘apple’ to other fruits based upon the similarity of shape and size. Another child may over-extend ‘apple’ to all round, red objects, based on similarity of shape and colour, yet ignoring the categories size and fruit.
Furthermore, phonology is another aspect of lexical development worth noting when discussing its predictability. Learning to produce the correct pronunciation is a required skill that happens simultaneously to word-referent association, but is very much a separate one. Research has been carried out into the patterns of phonological development that occur in infancy, with Jakobson’s 1968 research being a significant one. He analysed diary accounts and found that the same first sounds were recorded for infants between 12-18 months which encouraged Jakobson to suggest a “universal order of acquisition”. He findings included: first syllable structures are CV or CVCV reduplicated, labial sounds, most commonly [p] or [m], are the first consonants produced, with [t] then [k] following, and [a] is the first vowel sound produced, with [i] and [u] following (Ingram, 1976). This seems to suggest a predictability to children’s phonological development, although Ferguson and Garnica later suggested that Jakobson had failed to notice that [h] and [w] are often two of the earliest sounds learnt by infants (D, Ingram, 1976). The fact that this observation was not suggested originally by Jakobson suggests that there is more variation than he had believed, with the possibility that his list of the first acquired sounds of infants is not quite universal. Moreover, this questions the predictability of infants’ early speech sounds within their lexical development.
“Children often simply use a sound they know in place of one that they have not yet mastered” (McLean & McLean, 1999, p.179). However, this replacement tends not to be random- children usually swap the sound they cannot yet produce with a sound they can of similar quality, usually in terms of manner of articulation (McLean & McLean, 1999, p,179). My cousin demonstrated this with her pronunciation of my name Becca. She would say “Betta”- swapping the velar, plosive consonant [K] with the alveolar, plosive consonant [t]- as she had not yet mastered the [t] sound. Although less common, replacements of individual sounds can occur within manner of articulation. In multi-syllable words, it is common for young children to omit certain sounds that they are not familiar with making it less complex and simpler for them to attempt (McLean & McLean, 1999, p. 179). When my younger sister was an infant, she demonstrated omition of an individual sound when she attempted to produce the noun ‘crisps’. She could pronounce it as “crips”, omitting the alveolar fricative consonant [s] word-medially. The consonant cluster was too difficult for her to produce at the time- it was not that she could not pronounce the [s] sound as this was demonstrated at the end of the word, but rather that the [s]-[p]-[s] was too complex of a structure for her articulators to produce. The idea of children either replacing a sound with a similar one in terms if manner or place of articulation, or omitting a sound completely, or both, highly suggests that pronunciation is unpredictable. Research has provided us with the knowledge that these errors in a child’s early speech do occur, but which of these errors a child uses, how much of their vocabulary do they affect, and which specific words? We would not be able to predict this.
In conclusion, there are many aspects of lexical development that are somewhat predictable, for example, children’s first words being predominantly labels for objects and their related function words, children will make under- and over- extension errors early on in their semantic-lexical development, that infants will make phonetic errors in for form of substitutions and omissions. However, the specifics of these rather general ideas are much less predictable, for example: which objects and which related function words will a child first learn the labels for? Which concepts with children under-extend and which concepts will they over-extend? Which sounds will be substituted, and what sound will be the substitute? Which sounds will be omitted, within which words? Lexical development is a unique process for each child. Although broad aspects are predictable and certain patterns are assumed, individual differences will always occur.
References:
Barrett, M., Harris, M., & Chasin, J. (1991). Early lexical development and maternal speech: A comparison of children's initial and subsequent uses of words. Journal of Child Language, 18(1), 21–40. Available at: https://doi.org/10.1017/S0305000900013271 [Accessed: 13th Jan 2020]
Bloom, L. (1970). Language development; Form and Function in Emerging Grammars. Cambridge, Mass.: The M.I.T. Press.
Brooks, P. and Kempe, V. (2012). Language development. Chichester: BPS Blackwell.
Fernald, A., Perfors, A., & Marchman, V. A. (2006). Picking up speed in understanding: Speech processing efficiency and vocabulary growth across the 2nd year. Developmental Psychology. Available at: https://doi.org/10.1037/0012-1649.42.1.98 [Accessed: 13th Jan 2020]
Ingram, D. (1976) Phonological disability in children. London: Edward Arnold Ltd. McLean, J. and McLean, L. (1999). How children learn language. San Diego, Calif.: Singular Publishing Group.
Nelson, K. (1973). Structure and Strategy in Learning to Talk. Monographs of the Society for Research in Child Development, [online] . Available at: https://www.jstor.org/stable/1165788?seq=1. [Accessed: 13th Jan 2020]
Smith, L. and Yu, C. (2008). Infants rapidly learn word-referent mappings via cross-situational statistics. Cognition, [online] 106(3), pp.1558-1568. Available at: https://www.sciencedirect.com/science/article/abs/pii/S0010027707001795. [Accessed:13th Jan 2020]
Comments