This paper reports an experimental investigation into the acquisition of indirect object constructions by English children between the ages of 5;0 and 10;0. Two constructions were tested: one in which the indirect object precedes the direct object as in Give the dog a bone; the other in which the indirect object is a prepositional phrase starting with to and follows the direct object as in Give a bone to the dog. For the purpose of the following discussion the first construction will be labelled the order construction; the second construction the toconstruction. As yet little research has investigated the acquisition of these two constructions in English. Fraser, Bellugi & Brown (1963) found that the indirect object was the most difficult out of ten constructions that they tested with children with a mean age of 3;4. Carrow (1968) showed that 60% of children understood Show me the ball by at least 3;0 but that the same percentage understood She showed the girl the boy only at 4;6. Nelson (1973) reported that Give the — to your mommywas more easily understood by children aged 1;5 and 1;8 than Give me the —, though the syntactic difference is here complicated by the other differences between mommy and me. Stayton (1972), in unpublished work reported in Waryas & Stremel (1974), showed that children between 2;10 and 4;4 responded better to the to construction. Waryas & Stemel (1974) discovered an overwhelming preference among adults for the to construction. Finally Cromer (1975) showed that the order construction was understood by two-thirds of children only after the age of 8 ; 6. Existing research, therefore, while tending to show that all ages 'prefer' the to construction, tells us little about the child's comparative acquisition of the two structures between 5;0 and 10;0, and the following experiment was accordingly carried out. Ninety children acted as Subjects, fifteen at each year from 5;0 to 10;0. They were all attending primary schools in East London and had parents who were native speakers of English.
A list of eight sentences was constructed, all starting with give and using combinations of the girl, the man, the car, and the book as direct and indirect objects. Four sentences had the order construction, e.g. Give the girl the car; four the to construction, e.g. Give the book to the man. In addition two sentences had animate direct and indirect objects, two had inanimate direct and animate indirect, two had animate direct and inanimate indirect, and two had inanimate direct and indirect objects. These eight sentences were doubled by rotating the pairs of nouns used. Two sets of the resulting sixteen sentences were prepared in different randomized orders.
The children were tested individually. They were seated at a table on which four toys were displayed, a man doll, a car, a book, and a girl doll, always in the same order. Each child was told 'Give me the one I say' and was asked 'Give me the car' and so on, mentioning the four toys in random order. Then the child was told 'Now you've got to give them to each other not to me' and the test sentences were read. A record was kept of which toy the child moved and where he placed it. A few children initially moved two toys; they were told 'No, you must move only one toy', and after this further instruction all the children's responses were unambiguous.
Table 1. Correct answers on sentences with order constructions and to constructions (maximum 120 per cell)
Age |
5 |
6 |
7 |
8 |
9 |
10 |
order construction |
37 |
34 |
32 |
48 |
74 |
63 |
to construction |
106 |
111 |
117 |
120 |
119 |
120 |
In Table 1 the results are compared for sentences using the order construction and the to construction: at all ages from 5;0 to 10;0 children made many mistakes on the order construction and very few mistakes on the to construction (P < 0.01, sign test); there was greater accuracy with age for the order construction (P < 0.01, x2 test) but not significantly for the to construction. The results were also compared for sentences with the order construction that differed according to the animacy of the nouns; at all ages those sentences in which the direct object was inanimate and the indirect object animate were easier than those in which the direct object was animate and indirect object inanimate (all ages except 7;0, P < 0.01, sign test; 7;0 non-significant); at all ages sentences in which the direct object was animate and the indirect object inanimate were easier than those in which both objects were inanimate (all ages except 9;0, P < 0.03, sign test; 9;0 non-significant). These results show that English children between 5;0 and 10;0 understand the to construction before they understand the order construction; that in the case of the to construction there were no effects of the varying animacy of the nouns but that in the case of the order construction there were effects of animacy.
The results of this experiment throw an interesting light on several issues in child language. It has been claimed by Roeper (1973), and supported with evidence from German, that ‘humans are innately predisposed to have the indirect object precede the direct object'. The present results show that English children at this age pursue exactly the opposite strategy and prefer direct objects to precede indirect objects. McNeill, Yukawa & McNeill (1971) suggest an alternative strategy - that children prefer 'secondary' forms to be marked - and justify it from Japanese. If the to construction is considered 'marked' this would be supported by the present results, though reasons would have to be given for its 'markedness' that were other than post hoc. Cromer's results with artificial marking by postpositions do not, however, support McNeill et al.'s hypothesis (Cromer 1975). Macnamara (1972) also argues in favour of a cognitive basis for language learning by claiming that the child's decision to treat the two constructions tested here as stylistic variants' must depend on the child's appreciation that while books can be given to people, people cannot be given to books'. The present results show on the contrary that the child of five is as ready to give a man to a car as a car to a man, in the case of the to construction, and that he only uses a semantic strategy with the order construction. Finally these results support the preference revealed in other research for the to construction; if a criterion of explanatory adequacy is employed as described elsewhere (Cook 1974) they should assign the toconstruction priority over the order construction in a transformational generative grammar of English. This discussion should, however, be qualified by pointing out that the youngest children tested were 5 years old; it is possible that younger children still may adopt a different strategy for dealing with these two constructions.
REFERENCES
Carrow, M. A. (1968). The development of auditory comprehension of language structure in children. JSpHDis33. 99-111.
Cook, V. J. (1974). Is explanatory adequacy adequate? Linguistics33. 21-31.
Cromer, R. F. (1975). An experimental investigation of a putative linguistic universal: marking and the indirect object. JExpChPsych20. 73-80.
Fraser, C, Bellugi, U. & Brown, R. (1963). Control of grammar in imitation, comprehension, and production. JVLVB2. 121-35.
Macnamara, J. (1972). Cognitive basis of language learning in infants. PsychRev79. 1-13.
McNeill, D., Yukawa, R. & McNeill, N. B. (1971). The acquisition of direct and indirect objects in Japanese. ChDev42. 237-49.
Nelson, K. (1973). Structure and strategy in learning to talk. Monogr. Soc. Res. Ch. Devel. 38.
Roeper, T. (1973). Theoretical implications of word order, topicalization and inflections in German language acquisition. In C. Ferguson & D. Slobin (eds), Studies of language development. New York: Holt, Rinehart & Winston.
Stayton, B. (1972). The acquisition of direct and indirect objects in English. Unpublished manuscript, University of Kansas.
Waryas, C. & Stremel, K. (1974). On the preferred form of the double object construction. JPsycholingRes3. 271-80.