According to Lenneberg, there is a sequence of ‘milestones’ as language acquisition progresses and, within a Piagetian approach, these are correlated with age as well as cognitive development (sensorimotor stage, preoperational stage, concrete operational, etc). In other words, children acquire language in a highly structured way, making hypotheses and testing them against the linguistic input they are exposed to. Nevertheless, the claim—advocated by many experts (Pinker, 1994; Chomsky, et al.)—that a child acts more like a ‘grammatical genius’ than an empty slate, a ‘tabula rasa’ (Locke), ready to be written on and forged by experience, ‘seem[s] jarring to those who think of language as the zenith of the human intellect and who think of instincts as brute impulses that compel furry or feathered zombies to build a dam or up and fly south’ (Pinker, 1994: 20). Rather than widening the innateness hypothesis debate, in the present study we will content ourselves with outlining the stages normally occurring in the acquisition of morphology and syntax from the ‘one-word’ stage to about age 5. A great number of theories have been put forth to account for the mystery of language acquisition; however, we will only briefly mention one or two of these, as an in-depth analysis is not within the purview of this essay. What we can vouch for is that, no matter which theory one aligns oneself with, the acquisition of language is somewhat like the problem of reconstructing a dinosaur while the bones are still being excavated. It can happen that after you have connected what you earnestly believe are the hind legs you find that they are the jaw bones (McNeill, 1970: vii, cited in Aitchison, 1989: 109).
By the age of one year, the toddler’s speech comprises ‘one-word’ or ‘single-word’ utterances, such as BA, HA, QUA, DOGGIE, or ‘single-word approximations of frequently used adult phrases’, such as WASAT? (“what’s that?”) or ANKU (“thank you”) (Owens, 1996: 238). Undoubtedly, at this stage the child begins to develop the ability to convey meanings prior to the mastery of a systematic, consistent or coherent communicative system. As Scovel (1998: 13) notes,
[i]t is no wonder then that parents are excited by their child’s first word: it represents a step into symbolic communication, and it signifies the start of the rapid vocabulary growth with which thoughts, feelings, and perceptions, as well as other areas of linguistic development, are framed.
During this holophrastic period, one-word utterances seem to function as full-blown sentences, as is evidenced by three cues which the child’s speech is imbued with, so to speak: context, gesture, and intonation (see Taylor & Taylor, 1990: 255). For example, any one-word utterance may be taken to have multiple meanings, depending on the context. A French toddler said MAMAN (“mummy”) when he wanted to suckle; to be picked up by his father; or to draw attention to an object (Guillaume, 1927, cited in Taylor & Taylor, 1990: 256). Furthermore, a toddler’s speech may be “interspersed” with gestures, to the effect that she will probably raise her hands when she wants to be picked up, or she will say “bye-bye” with a waving hand. According to Taylor & Taylor (1990: 256), ‘it is the gesture that conveys the different communicative functions of assertion, request, protest, or denial, while the word names the object’—a view in keeping with Owens’ claim that ‘[f]irst words fill a variety of illocutionary functions, such as answering, requesting, declaring, and seeking information’ (Owens, 1996: 239). Finally, toddlers may vary their intonation in uttering a word, such as mama, either raising it to mean “Where have you been, mama?” or dropping it to mean “There you are, mama” (Taylor & Taylor, 1990: 256).
But interpreting a toddler’s one-word utterances is not as straightforward as many people would think. In fact, the task of extracting meaning from the morass of associations which a child makes between a single-word utterance and a referent in his environment is fraught with difficulties, as we shall see. Suppose that a child says BA when he is having a bath; when holding an empty mug; and also when he seems to be referring to the kitchen taps (see Aitchison, 1989: 109). How is one to account for the child’s tendency? Four possibilities have been put forward, which we will briefly discuss.
The first one is that the child engages in the naming of objects in order to prove that he knows them but, in doing so, he has overgeneralised the word BA. More specifically, he knows that BA means “bath” but labours under the misconception that BA can apply to anything containing liquid. A second, and perhaps more plausible explanation, has been proposed by Vygotsky. According to him, children focus attention on one aspect of an object, which can lead to confusing overgeneralisations. For instance, a child said QUA to refer to a duck, milk, a coin, and even a teddy bear’s eye (Vygotsky, 1962: 70, cited in Aitchison, 1989: 110). Since QUA originally referred to a duck on a pond, the child focused attention on the liquid element, i.e., water, and generalised it to milk. However, QUA was also used to refer to a coin with an eagle resembling the duck on the pond. Then, the child used QUA to refer to any round object, such as a teddy bear’s eye. This ‘chain complex’ phenomenon must also apply in the case of BA, whereby the child, focusing attention on the liquid element, overgeneralised the word to milk, on the one hand, and on the other, remembering the bath taps, overgeneralised BA to the kitchen taps.
Yet, David McNeill, a psychologist at the University of Chicago, takes a different view. For him, one-word utterances exhibit a degree of linguistic sophistication which goes far beyond the sound spoken. The child is not merely stringing together sound sequences to name objects, but is actually uttering holophrases, as has been mentioned above. In this light, BA can mean “I am taking a bath” or “My toy has fallen in the bath.” Moreover, when the child utters a word at this stage, it is presumably because he shows an understanding of spatial and other relations: “There are taps like this in the bathroom” or “We put the bananas on the fridge.” Such a claim, though, has been dismissed by many as over-ambitious and imaginative. Lois Bloom has a different view.
After carefully analysing the single-word utterances that her daughter produced, she has come to the conclusion that the meaning of such utterances differs according to the age of the child. Thus, when her daughter said MUMMY at the age of sixteen months, she might mean “That’s Mummy,” whereas three months later she must have been trying to posit some kind of interaction between her mother and the environment, as when she pointed to a cup and said MUMMY.
A great many linguists and psychologists have interested themselves in analysing one-word utterances and, particularly, what the child is trying to do with them. If a child says GA, is she naming the cat; is she actually asking for the cat; or is she complaining because the cat is outside and she wants her parents to let it in? As is obvious, there is no straightforward explanation for these confusing data. ‘All researchers, to some extent, see what they want to see’ (Aitchison, 1989: 113).
Before moving on to the next stage, that is, two-word utterances, it is of consequence to note that the views of most researchers seem to converge on the following points: During the one-word stage, the child shows signs of communicating facts about her environment. Inasmuch as she lacks the mature communicative system of adults, she depends on gestures to convey intent, coupled with phonetic approximations of adult words, such as BA, HA, QUA. ‘At first, functions are isomorphic; that is, specific words…are used with each function. Gradually, functions fulfilled by words emerge. As words and functions increase, words and utterances become more flexible and multifunctional’ (Owens, 1996: 241). In this sense, toddlers’ utterances can be said to perform various speech acts, called primitive speech acts (PSAs) (Dore, 1974, cited in Owens, 1996: 241). What is more, the common denominator in language acquisition at this stage is children’s ability to ‘understand many more words than they can produce’ (Taylor & Taylor, 1990: 257).
By the age of eighteen months or later, toddlers’ speech is characterised by the use of two (or sometimes three) words strung together in an utterance, as in MORE DOLL and NO UP. Braine (1976, cited in Taylor & Taylor, 1990: 292) noted three types of patterns that children’s speech seems to follow at this two-word stage. In what he dubbed a groping pattern, the toddler is actually groping to impart some kind of meaning in the absence of any knowledge of a consistent set of syntactic rules. Thus, EAT COOKIE or COOKIE EAT are possible utterances, whereby the kid strings two words together meaningfully, ‘without regard for word order or position’ (Owens, 1996: 261). In the positional-associative pattern, the position of words is consistent without there being a consistent underlying rule. It may be that the child is repeating patterns found in adult speech, such as STOP THAT or COME HERE (ibid.: 261). Finally, in the positional-productive pattern, words combine in a consistent and productive way, as in EAT COOKIE, THROW BALL, DRINK JUICE, DADDY SHOE, and so on, which attests to the application of certain rules such as action + object or Possessor + possession (ibid.: 261-262).
By examining sixteen corpora of two-word combinations, Braine came to the conclusion that there must be a grammar for two-word utterances, which he called pivot grammar. A pivot word is a word occurring constantly, while an open-class word is used inconsistently. For example, a sentence can be decomposed into a pivot word and an open word, as in SEE DOGGIE (S P1 + O). In the toddler’s speech, there are numerous possible constructions such as BOOT OFF (O + P2) or COOKIE EAT (O + P3) but not SEE OFF (P1 + P2) (see Taylor & Taylor, 1990: 293). Braine thought that he had come across a universal grammar. Nevertheless, not every toddler’s speech seemed to follow these rules. After all, how can one characterise a sentence such as DADDY SHOE as O + O? Such a rule cannot distinguish between several possible interpretations: “Daddy is putting on his shoe,” “That’s Daddy’s shoe,” or “Daddy is looking for his shoe.” As Aitchison (1989: 116) notes, ‘[pivot grammars] only describe the rules used by a small number of children—or, perhaps, more accurately, they characterize only a small portion of the output of most children’.
A pivot grammar may have been proved lacking in some respects, yet it has helped shed some light on children’s ability to form and apply rules, as was intimated above. Brown (1973: 173, cited in Aitchison, 1989: 119) has suggested eight minimal two-term relations, along with three basic operations of reference, in an attempt to capture the concepts possibly expressed at the two-word stage (see the chart below, cited in Aitchison, 1989: 119).
1. Agent action MUMMY PUSH
2. Action and object EAT DINNER
3. Agent and object MUMMY PIGTAIL
4. Action and location PLAY GARDEN
5. Entity and location COOKIE PLATE
6. Possessor and possession MUMMY SCARF
7. Attributive and entity GREEN CAR
8. Demonstrative and entity THAT BUTTERFLY
9. Nomination THIS (IS A) TRUCK
10. Recurrence MORE MILK
11. Non-existence ALLGONE EGG
Certainly, this ‘highly consistent word order makes it unlikely that the sequences are random juxtapositions’ (Aitchison, 1989: 117); but how much leeway is one left with in holding that children at this stage exhibit grammatical knowledge? Chomsky’s claim that children are equipped with a Language Acquisition Device (LAD), albeit plausible and illuminating, has sparked considerable controversy, and many people have shifted away from it (we will briefly discuss Chomsky’s LAD later on). As a matter of fact, rather than showing evidence of grammatical knowledge, these two-word utterances ‘merely reveal an awareness that meaning relationships need to be expressed consistently’ (Aitchison, 1989: 112). Furthermore, one could assert that ‘[t]he case categories of actor and action, cued partly by the meanings of words, may be easier to acquire than the grammatical categories of subject and verb’ (Taylor & Taylor, 1990: 294).
This, however, leaves us with a problem. As Aitchison (1989: 120) notes,
If we assume that two-word utterances show grammatical knowledge (which would be fanciful) then we have to specify exactly what kind of grammar we are dealing with. If, on the other hand, we do not regard them as showing evidence of grammar, then we have to find out when children start having a primitive syntax. In this case, we have to assume that language learning is a discontinuous process. Children start with one kind of system, and then shift over to another, syntactic one. We may be dealing with a tadpole-to-frog phenomenon (Gleitman and Wanner 1982), in which the immature tadpole behaves rather differently from the mature frog.
A possible solution to this predicament is to resume our discussion, in terms of trying to figure out the ways in which children acquire grammatical and lexical morphemes, as well as questions, negatives, and other constructions.
Moving away from the two-word stage, or telegraphic speech, of toddlers, where utterances lack most grammatical and lexical morphemes, children at the age of two to two and a half years gradually introduce some inflections. What is most interesting is the order of acquisition. For example, children all over the world, regardless of their mother tongue, seem to acquire first the progressive –ing, then the spatial prepositions on, under, etc., then the plural –s, the irregular past, the possessive ’s, the copula am, is, are, and the articles a and the. At the age of 3, the child will acquire the regular past –ed, the 3rd person singular –s, the full progressive am, is, are + -ing, the shortened copula, and the shortened progressive (see Aitchison, 1989: 75; Owens, 1996: 305-341; Taylor & Taylor, 1990: 296). Let us briefly examine each stage.
Children express the present progressive by introducing the grammatical morpheme –ing, while omitting the auxiliary to be, as in MARY SINGING, I EATING. Moreover, they seem to know that the progressive can be used with action verbs but not with verbs of state (need, know, like, etc.). ‘The child probably learns the rule one verb at a time by applying it to individual verbs to determine whether they are “ingable” (R. Brown, 1973)’ (Owens, 1996: 305-306). Later come the prepositions in, out, away, under, over, with in and on looming largest in everyday speech. At this stage, the child knows that we put something in a cup, or something is on the table, but she cannot comprehend the meaning of behind, between, or in front of (ibid.: 307).
In general, the regular plural appears in short phrases and then in long sentences. According to Miller and Ervin-Tripp (1964, cited in Owens, 1996: 307), there are four phases of plural acquisition. First, the child does not distinguish between singular and plural; he just uses more to mark the plural, as in MORE DOGGIE. Second, the plural marker is used on plural words that the child is frequently exposed to. Next, he generalises plural to words such as foot and mouse, which become foots and mouses. Finally, the regular and irregular plural are mastered and used appropriately.
Interestingly, children acquire the irregular past earlier than the regular past, with such irregular verbs as ate, wrote, and drank frequently used. However, despite the belief that practice makes perfect, later on children use the regular past –ed with many irregular verbs, thus coming up with goed, comed, and falled. Pertaining to the possessive ’s, it is initially attached to animate nouns, such as Mommy or doggie, to form Mommy’s or doggie’s, and then to inanimate nouns, such as table or leg, to form table’s or leg’s.
The copula am, is, are, as in “I am a teacher,” is acquired first in full or uncontractible form, probably because the child hears adult sentences such as “I am” in response to the question, “Who is hungry?”, where it is impossible to contract the copula (“I’m”). Apropos of the articles a and the, for adults, the indefinite article a is used for non-specific reference, and the definite the denotes specific reference. For children, the article a initially predominates, while later on the is overused. In general, though, articles are first acquired in a nominative situation, such as “This is a cat” or “Where is the doggie?”. Furthermore, children at about four years of age know that a can be used with countable nouns but not with uncountable or mass nouns.
Later comes the regular past –ed, the acquisition of which leads to overgeneralisations such as comed, eated, or drinked, as we saw above. With regard to the 3rd person singular –s, children seem to be late in mastering it, especially with irregular verbs such as do and have. The full progressive am, is, are + -ing is acquired at about three and a half to four years of age. For example, I AM SINGING, HE IS EATING. Then, the shortened auxiliary appears: I’M SINGING, HE’S EATING.
A similar consistency of order applies to the acquisition of questions and negatives. Indubitably, children have plenty of opportunities to hear questions, and initially they answer wh-questions with yes or no, which is incorrect. For example, an adult may ask, “Why is the boy crying?” and receive no as an answer. In addition, children at the age of 3 usually mistake why questions for what questions, as in the example:
ADULT: Why is Mummy eating?
Besides, children make mistakes in the production of questions, going through three stages: First, the wh-word goes in front of the rest of the sentence.
WHAT DADDY DOING?
WHY SARAH CRYING?
WHERE YOU GO?
Later on, the child adds the auxiliary verb, which follows the subject.
WHERE HE WILL GO?
WHAT HE IS DOING?
WHY SARAH IS CRYING?
Then, she produces correct sentences such as:
WHAT IS DADDY DOING?
WHERE WILL YOU GO?
WHY DO YOU WORK?
As Aitchison (1989: 82-83) observes, ‘all English children tend to follow the same pattern…[and] the actual age at which each stage is reached is irrelevant. It is the order which matters’.
As far as negatives are concerned, we see that their acquisition, too, follows three stages: First, the child places NO or NOT in front of the sentence to produce:
NO WANT EAT
NO MUMMY GO
NO GO BED
This rule, however, which children seem to have formed, is soon replaced by a new one that says, “Put NO or NOT after the first noun phrase,” as in
HE NO DRINK WATER
I DON’T GO BED
THIS NO RED
I CAN’T WALK
At this stage, i.e., at the age of two and a half years, the child regards DON’T and CAN’T as alternatives to NO and NOT, for he has not realised that they consist of two elements, DO + NOT and CAN + NOT. He seems to know, though, that CAN’T and DON’T are not used with verbs ending in –ing, so he never says *I CAN’T EATING or *I DON’T SLEEPING.
Later, the child realises that CAN’T and DON’T contain two separate words and produces:
SARAH CAN N’T HAVE ONE
I AM NOT SICK
HE WAS NOT TALL
YOU DID N’T CAUGHT ME
Finally, the child amends sentences such as YOU DIDN’T CAUGHT ME and I DIDN’T SPILLED IT to YOU DIDN’T CATCH ME and I DIDN’T SPILL IT, respectively.
Other complex constructions that 2- to 3-year-olds produce are the following:
I WANT DADDY DO THAT
I DON’T WANT YOU TAKE THAT
According to Wells (1985a, cited in Taylor & Taylor, 1990: 318-319), complex constructions at this stage evince three characteristics: a) The embedded clause or complement (DADDY DO THAT) could stand on its own as an independent sentence; b) the complementizer that is absent; c) the object noun, but not the subject noun, is expanded into a phrase or clause (there are no sentences such as “That he came made me happy”).
At the age of 5, the child can produce complex sentences such as:
I’M SORRY TO SAY THAT THE PEOPLE THAT MADE IT, I THINK, ARE A BIT SILLY (from Taylor & Taylor, 1990: 319).
Remarkably, a five-year-old gives the impression of having mastered his mother tongue. Yet, this is not the case. A group of 5- to 8-year-olds were shown a blindfolded doll, and were asked: “Is this doll hard to see or easy to see?” All the 5- and 6-year-olds said HARD TO SEE, and so did some of the 7- and 8-year-olds. This shows that children of 5 and 6 do not realise that constructions such as HARD TO SEE and EASY TO SEE have distinct underlying meanings.
A great many theories have been put forward to account for the mystery of language acquisition, but we will only briefly mention two of them: Chomsky’s content approach and an alternative process approach (see Aitchison, 1989, chapter 7 for details). According to the content approach, a child is endowed with language-specific knowledge, a Language Acquisition Device (LAD), which is independent of general intelligence. To corroborate his theory, Chomsky has tried to show that children know in advance certain facts about language, such as surface and deep structure; they are intuitively aware of universal constraints; and they set ‘switches’, in order to cope with the specific language they are learning. Chomsky’s theory has been rather influential, yet it has been disputed, as many of its tenets are not borne out by the evidence.
Within the process approach, there are two theories. The first one views the child as a general problem solver who, because of her needs, her general cognitive development, and the parental output she is exposed to, is “goaded into linguistic action,” as it were. The second one regards the child, not as a general problem-solver, but as one specific to language. In short, it does not presuppose the existence of innate linguistic knowledge, but rather it sees the child as an experimenter who is trying to solve linguistic puzzles. Like the content approach, the process approach is unable to account for language acquisition. Unfortunately, it is outside the remit of this study to consider in depth the putative conundrums of these theories.
By way of a conclusion, we could summarise the main points of our discussion. All over the world, children acquire language in much the same way, following a highly structured and sequenced set of milestones. At the age of 2, toddlers everywhere are expected to have mastered one-word utterances and to produce two-word combinations, which ‘encode several semantic categories (e.g., actor-action) and speech acts (e.g., request and rejection)’ (Taylor & Taylor, 1990: 323). Grammatical morphemes are initially absent but are gradually mastered over a period of years. The acquisition of wh-questions follows three stages, whereby the wh-word first goes in front of the sentence, to be mastered a few months later. At the age of two to two and a half, complex sentences lack some features, such as a complementizer, but by the age of 3 onwards, they are correctly formed, and the child’s language resembles adult speech. Taylor & Taylor (1990: 325) succinctly summarise the main elements that undergird language acquisition:
Language is acquired under three conditions: the rapid and sequential maturation and cognitive development; the pragmatic need to exchange information and to control others’ actions; and an exposure to appropriate speech input. These three conditions may be universal, and the first two may be innate as well.
- Aitchison, J. 1989. The Articulate Mammal. London: Routledge.
- Owens, R. 1996. Language Development. USA: Macmillan.
- Pinker, S. 1994. The Language Instinct. London: Penguin.
- Scovel, S. 1998. Psycholinguistics. Oxford: OUP.
- Taylor, I. and Taylor, M. 1990. Psycholinguistics. Learning and Using Language. USA: Prentice Hall.
Source: Dr. Dimitris Thanasoulas