What Do You Mean, No? Toddlers’ Comprehension of Logical “No” and “Not”

For adults, “no” and “not” change the truth-value of sentences they compose with. To investigate children’s emerging understanding of these words, an experimenter hid a ball in a bucket or a truck, then gave an affirmative or negative clue (Experiment 1: “It’s not in the bucket”; Experiment 2: “Is it in the bucket?”; “No, it’s not”). Replicating Austin, Theakston, Lieven, & Tomasello (2014), children only understood logical “no” and “not” after age two, long after they say “no” but around the time they say “not” and use both words to deny statements. To investigate whether this simply reflects improving inhibitory control, in Experiment 3 we showed children that one container did or did not hold the ball. Twentymonth-olds now succeeded. We discuss two possible factors limiting learning both “no” and “not”—a purely linguistic difficulty learning the labels, and the possibility that negation is unavailable to thought before age two. One of the most remarkable things about human cognition is our ability to take a finite number of concepts and combine them to generate an infinite number of meaningful thoughts. An English speaker who has never heard the sentence, “There are no bears on Mars,” has no trouble understanding what it means. Furthermore, we are able to judge that it is very likely true, and can infer that, if it is, then there are also no brown bears on Mars, no bear cubs, and no bears climbing Martian trees. The ease with which we can understand new thoughts, make judgments about their truth, and reason about related thoughts has just one plausible explanation—complex thoughts are generated by a system for combining the meanings of concepts in systematic ways. Negation is an operation that flips the truth-value of a sentence, making negation words like “no” and ”not” function words with highly abstract meanings. Unlike the vast majority of content words—nouns, verbs, and adjectives—they do not refer to anything in the world. Their meaning comes from the systematic way that they alter the meanings of the propositions they combine with. Yet despite the abstract meaning of negation, there is not a language in the world without a word for it and not a culture that has ever been reported to dispense with it in their thought. This article explores the developmental origins of negation in language, taking it as a case study of the development of abstract, combinatorial concepts. By looking at the timing of children’s comprehension of “no” and “not”, we can learn about the conceptual and linguistic abilities that are required for using truth-functional negation. The age at which children first understand the truthfunctional meaning of a negation word provides an upper bound on when truth-functional negation is available to productive thought. The relative timing of learning “no” and “not” is also informative. If children make the mappings for both “no” and “not” at the same time, it would suggest that a single conceptual or linguistic factor—something that the words have in common—limits the CONTACT Roman Feiman [email protected] Department of Psychology, University of California, San Diego, 3316 McGill Hall, 9500 Gilman Dr., La Jolla, CA 92093. Color versions of one or more of the figures in the article can be found online at www.tandfonline.com/hlld. © 2017 Taylor & Francis Group, LLC LANGUAGE LEARNING AND DEVELOPMENT https://doi.org/10.1080/15475441.2017.1317253 mapping of both words. If, on the other hand, children map truth-functional negation to “no” and “not” at different times, the mapping of the words is likely constrained by different factors—for example, by the input frequency of each word or their different grammatical properties. What do we know about the acquisition of the truth-functional meanings of “no” and “not”? A first step is to look at when these words enter children’s productive vocabulary. While over half of 15-month-olds say the word “no”, it is not until 27 months—a full year later—that half the children say “not” (Dale & Fenson, 1996; see Figure 1). However, 15-month-olds may not be expressing truth-functional negation when they say “no”. When children first begin to produce negation words, they use them for a narrower set of functions than do adults—to reject offers or commands, or to comment on the nonexistence of something they expected (Bloom, 1970; Pea, 1980a). It is not until 24–30 months that children use “no” or “not” to deny the truth of others’ utterances (Bloom, 1970; Choi, 1988; Hummer, Wimmer, & Antes, 1993; Pea, 1980b, 1982), as when a parent points to a dog and asks, “Is this a bear?”, and the child answers, “no” or “not bear”. This denial function is the latest to emerge in children’s usage, and is most closely associated with truth-functional negation (Pea, 1982). The production of denial negations tracks closely with the emergence of “not” in a child’s productive vocabulary (Cameron-Faulkner, Lieven, & Theakston, 2007; Choi, 1988). Given that children begin to produce denials between 24 and 30 months, why not simply conclude that this is when children map truth-functional negation to both “no” and “not”? The difficulty with data from language production is that it is impossible to distinguish the full range of meanings that a child might know from their pattern of usage. Perhaps a young child is in a position where she knows little about the world, but is confident about her distaste for bathing, and therefore has less occasion to deny a factual claim than to reject the offer of bath time, even if she can use negation as a truth-functional operator. To address this ambiguity, we can examine children’s comprehension of negation. If negation words do not have a truth-functional meaning for young children, they should not understand sentences where negation words invert truth values. Most studies of the comprehension of negation indicate that understanding truth-functional negation poses a significant challenge for children, even well past the age at which they are Figure 1. Production of the words “no” and “not” by age, based on data from Macarthur-Bates CDI population norms. Children’s age in months is along the X-axis, and the percent of children producing the words is on the Y-axis. Points are means at each month. Lines are logistic regression curves. This figure was generated from the Wordbank repository (Frank, Braginsky, Yurovsky, & Marchman, 2016). 2 R. FEIMAN ET AL.