Language acquisition and conceptual development: Children's weak interpretations of universally quantified questions
Language acquisition and conceptual development: Children's weak interpretations of universally quantified questions
复制标题
语言习得和概念发展：儿童对普遍量化问题的薄弱解释
DOI：
10.1017/cbo9780511620669.014
复制
发表时间：
2001
影响因子：
1.2
通讯作者：
K. Drozd
中科院分区：
文献类型：
--
作者：
K. Drozd
自皮亚杰的班级包容研究以来，儿童用像Every和All这样的通用量词进行推理一直是发展心理学和心理语言学的一个重要课题。一个谜题是，当被问到像是不是每个男孩都骑大象这样的问题时，孩子们为什么会做出两种特殊的出乎意料的反应？当孩子们看到一张如图12.1a和b所示的图片，并被要求判断是否每个男孩都在那个背景下骑大象时，就会出现一种错误。图12.1a中的图片描述了一个不完整的一对一匹配上下文，其中三个男孩每个人骑着一头不同的大象，留下一头大象没有骑。图12.1b描绘了一个多对一的匹配，其中三个男孩都骑着一头大象，剩下两头大象没有骑。成年人通常会在这些任务上说是，但有些孩子会说不是。当被要求解释时，他们会指出没有骑过的一头或几头大象，这表明他们希望看到每头大象都与一个男孩配对。让我们称其为“彻底配对错误”。另一种类型的错误发生在向孩子们展示如图12.2所示的图片并询问相同的问题时。一些孩子在这项任务上的回答是肯定的，这表明他们可能无法在整个物件领域进行详尽的搜索，即使该领域在感知上是可用的。让我们称其为“不完全配对错误”。为什么孩子们会犯这些错误？以往的研究都假设儿童知道全称量词的含义，但为普遍量化的句子构建了句法或语义表征，这并没有适当地限制量词的范围是如何选择的。正如我将要展示的，这些描述都不能解释这些错误所特有的依赖于语境的特定类型的解释。相反，我将支持另一种假设，即儿童将全称量词分析为弱量词。弱量词的类别包括两个、一些和许多，它们表现出特定的上下文依赖和推理属性，这些属性将它们与每个、所有和大多数等所谓的强量词区分开来。我支持弱量词假设(1)，通过展示两个错误孩子的属性
Children's reasoning with universal quantifiers like every and all has been an important topic in developmental psychology and psycholinguistics since Piaget's class inclusion studies. One puzzle is why children make two particular kinds of unexpected responses when asked questions like Is every boy riding an elephant?One type of error occurs when a child is presented with a picture like those shown in figures 12.1 a and b and asked to judge whether every boy is riding an elephant in that context. The picture in figure 12.1 a depicts an incomplete one-to-one matching context in which each of three boys is riding a different elephant, leaving one elephant unridden. Figure 12.1 b depicts a many-toone matching in which three boys are all riding one elephant, leaving two elephants unridden. Adults typically say yes on these tasks, but some children say no. When asked to explain, they point out the unridden elephant or elephants, in a way that suggests that they expected to see each of the elephants matched with a boy. Let us call this the “exhaustive pairing error." The other type of error occurs when children are presented with a picture like that in figure 12.2 and asked the same question. Some children answer yes on this task, suggesting that they may fail to search exhaustively through an entire domain of objects, even when that domain is perceptually available. Let us call this the" underexhaustive pairing error." Why do children make these errors? Previous accounts all assume that children know the meaning of universal quantifiers but construct syntactic or semantic representations for universally quantified sentences which do not properly constrain how the domain of quantification is selected. As I will show, none of these accounts can explain the specific kinds of contextdependent interpretations which characterize these errors. I will argue instead for an alternative hypothesis, namely that children analyze universal quantifiers as weak quantifiers. The class of weak quantifiers, which includes two, some, and many, exhibits particular contextual dependencies and inferential properties which distinguish them from so-called" strong" quantifiers like every, all, and most. I support the Weak Quantifier Hypothesis (1) by showing how properties of both of the errors children