Acquisition of Quantification. Errors in Child Language

Table of Contents

1. Introduction

2. Quantifier Sentences and Prerequisites
2.1 What Do Children Know?
2.2 Errors in Child Language

3. The Symmetrical Account
3.1 Problems With The Symmetrical Account

4. Experience-based Approach

5. Felicity and Plausibility

6. Conclusion

7. Bibliography

1. Introduction

The acquisition of quantifiers has always been a widely discussed topic in linguistics. Do children really have access to an adult interpretation from birth on or do they need to learn it throughout their childhood? The main point hereby is that children seem to make a lot of mistakes while answering questions about quantification. They don’t seem to understand the meaning of every and all. These two quantifiers will be in the focus of this paper as well, because they are the most difficult ones and the ones where children make the most errors.

But what are these errors? And where do they come from? Some argue that children simply prefer a different interpretation than adults do or that they just have an inaccurate knowledge of the adult version. Others argue that children cannot distinguish part-whole relationships or that the position of the quantifier in the sentence is of importance. It could also be a parametric variation across languages (cf. Musolino et al. 2000). Often, linguists blame a wrong syntax to semantics mapping as well (cf. Geurts 2003).

The three most logical explanations for children’s errors with quantifier sentences are the symmetrical account, the experience-based approach and the felicity and pragmatic condition. The symmetrical account tries to explain the children’s errors linguistically by stating that children see objects in a symmetrical way. The experience-based approach considers the input that children get, and the felicity condition states that children only answer to plausible questions within plausible circumstances. My main question will be if children’s non-adult answers are due to linguistic factors (symmetrical account) or due to non-linguistic factors (experience-based, felicity). To examine this question I will first talk about quantifier sentences in general and what children need in order to interpret them. Then I will have a look on what children already know, before I will list all the errors they make in their interpretations. Furthermore, I will examine three different approaches concerning children’s errors in quantifier sentences: the symmetrical account, the experience-based approach and the felicity and pragmatic conditions. In the case of the symmetrical account, I will also talk about problems that occurred during experiments. After that I will draw a conclusion based on my research.

2. Quantifier Sentences and Prerequisites

To explain what quantifiers are, we need to differentiate between quantificational determiners like every and all and adverbs of quantification like always (cf. Guasti 2002: 88). In this paper, I will talk exclusively about determiners. One example of a quantifier sentence would be the following:

(1) Every farmer is feeding a donkey (Crain et al. 1996: 88).

In this sentence, every has wider scope than a, which is called the universal wide scope reading (cf. Guasti 2002: 322).

(2) A farmer is feeding every donkey (Crain et al. 1996: 88).

In the second sentence, a ranges over every. This is called the existential wide scope reading (cf. Guasti 2002: 323). To process this information, children need to be able to raise the quantifier from the VP to the IP position (Quantifier Raising – QR), as it can be seen in Figure 1. The quantifier is involved in a NP, it never stands alone. In (1), the NP is every farmer. This NP needs to be raised to IP. In contrast to QR, a quantifier sentence can also have a normal, linear order, like the one in (1). Here, children need to be able to understand both QR and the linear order (cf. Guasti 2002: 827). They furthermore need to understand that a quantificational sentence is a three place relation; in (1) for example, there is the farmer, the event of feeding, and a donkey. The quantifier is therefore referential (cf. Guasti 2002: 315). Children also need to understand the mapping from syntax to semantics.

Other typical test sentences are

(3) Every horse didn’t jump over the fence ( Musolino et al. 2006: 823)

And

(4) The Smurf didn’t buy every orange (Musolino et al. 2006: 823).

Of course, sentences like these, and most of the quantifier sentences, are ambiguous. Sentence (3) can mean that no horse jumped over the fence or that not all the horses jumped, but some did. Musolino et al. (2000) differentiated the two meanings in the

Abbildung in dieser Leseprobe nicht enthalten

Figure 1 Quantifier Raising (Guasti 2002: 315).

following way: In the first meaning, where no horse jumped, every>not. This is the isomorphic interpretation. In the second meaning, where not all the horses jumped, not>every. This is called the non-isomorphic interpretation. Children need to be aware of this ambiguity, but, somehow, they seem to prefer the isomorphic interpretation, whereas adults prefer the non-isomorphic one (cf. Musolino et al. 2000: 13). Sentences like (3) and (4) are also used in the so called Truth Value Judgment Test by Crain et al. (1996). This test usually involves a puppet recalling a story that happened before and the children have to say whether what the puppet says is true or false (cf. Guasti 2002: 320).

2.1 What Do Children Know?

Children basically have adult-like knowledge of quantifiers. They sometimes just prefer a different interpretation than adults do. According to Guasti (2002), children are able to distinguish between quantified and referential NPs. They are furthermore able to raise the quantifier to the IP position (QR) and they know the function of quantifiers (cf. 328).

According to Musolino et al. (2000), the difference between adults’ and children’s interpretation happens within certain boundaries though. These boundaries are derived from the theory of Universal Grammar (UG). Musolino et al. therefore conclude that children do have adult-like knowledge, but that their knowledge is incomplete in contrast to the adults’ knowledge (cf. 1). This fits the point in the first paragraph, where it is stated that children prefer different interpretations than adults do (also cf. Crain et al. 1996: 96). Children always interpret quantifier sentences as if the quantifier ranges over an event rather than an individual. Additionally, children only seem to have knowledge of the isomorphic interpretation of quantifier sentences (cf. Musolino et al. 2000: 13). Crain et al. (1996) also conclude that it could be that the quantifier itself is the problem, because children perform significantly better when there are two plural NPs instead of the universal quantifier in a transitive sentence. They also perform well in sentences like (5) and (6).

(5) Is every farmer a donkey-feeder? (Crain et al. 1996: 99).

(6) Every horse jumped over the log but every horse didn’t jump over the fence (Musolino et al. 2006: 825).

So, according to Musolino et al. (2006), 5-year-olds basically have adult-like knowledge, at least regarding (5) and (6). This includes syntactic and semantic knowledge and knowledge of covert displacement. The latter means that children are able to access the non-isomorphic interpretation of quantifier sentences because they are able to change positions within a sentence. This contradicts Musolino et al.’s findings in 2000, where they stated that children only have access to the isomorphic interpretation. Therefore, children also have the ability to differentiate between none and not all. Musolino et al. furthermore support the point that children simply prefer different interpretations than adults do (cf. 840 f.). They also point out that children have a bias to say yes rather than no (cf. 847).

[...]