The verb is the focal point of a sentence. It determines both the argument structure of a sentence and the semantic roles of its arguments. Learning the meaning of verbs in one's language is key to knowing the grammar itself. Yet verb learning is a challenge for children. This edited volume expands on the many factors that contribute to the intricacy of verb learning, bringing together an intriguing set of findings from many of the prominent researchers in the field.
What are the challenges involved in learning a verb? To correctly establish the meaning of a verb the child needs to find the appropriate underlying concept. Even if children have rich conceptual knowledge, finding the right concept is not easy. The concepts that verbs refer to are abstract and hard to pick out from the physical environment. Even highly imageable verbs like running differ greatly depending on who is performing the action; for instance, when grandma is running it looks very different than when a marathon champion does. Moreover, verbs don't label events directly, rather they refer to a perspective the speaker takes on an event. Finally, languages vary in how they group concepts under one verb meaning.
Even before children attempt to learn the meaning of a verb, they need to determine which words are the verbs. Part 1 of the book describes issues in isolating and identifying verbs from the speech stream. Part 2 then goes on to ask whether children possess rich enough conceptual knowledge to allow them to grasp verb meanings. The main conclusion from Part 2 is that children are equipped with suitable concepts. Nevertheless, verb learning is still a challenge as compared with noun learning. Parts 3 and 4 bring many insights into why that is, taking into account cross-linguistic and cross-cultural findings. Several chapters in Parts 3 and 4 lay out how the verb learning process might unfold and, in particular, suggest that it is guided by syntactic knowledge.
Words in the speech stream are not separated by silent pauses. Thus, the first step in learning any word is to segment it from the speech stream. Even this initial step is a challenge when it comes to verbs. Chapters 2 and 3 argue that it is easier to segment nouns than verbs. In Chapter 2, Nazzi and Houston find that 7·5-month-olds can segment nouns from fluent speech whereas only 13·5-month-olds succeed with verbs. They propose various prosodic factors that influence segmentation. In Chapter 3, Christiansen and Monaghan suggest that phonological cues facilitate verb segmentation in English more than distributional cues.
Once a word has been identified, how would a child know to classify it as a verb? After all, meaning cannot be the sole determining factor since many verbs do not refer to actions, and many nouns do. In Chapter 1, Mintz proposes that children can use a kind of distributional information, which he terms ‘frequent frames’, to categorize verbs. The frames are determined by taking the first and last words from three word sequences in sentences. The middle words in the sequence belong to the class that the frame defines. For example, ‘you _ it’ is one frame that occurs frequently for the class of verbs. Based on corpus data from child-directed speech, Mintz shows that these frames appropriately break up nouns and verbs into different sets. In addition, he presents evidence that 12-month-olds can use distributional information of this sort to categorize novel words.
Part 2 of the book deals with another prerequisite for verb learning: children's grasp of the underlying concepts. The concepts that verbs refer to are relational and abstract. Are verbs hard to learn because children do not yet possess relational concepts? In Chapter 4, Mandler maintains that this is not the case. Rather, by the time infants begin to speak their first words, they can reason about many kinds of relational events. For example, infants differentiate between causal and non-causal actions, and make sense of objects moving into and out of containers. These concepts, however, do not map directly onto verbs. For example, the actions of putting a piece into a puzzle, and an apple in a bowl, are described in English using the same verb put in, but by two different verbs in Korean. Korean distinguishes those two actions on the basis of tightness of fit. The piece fits tightly in the puzzle, but the apple does not fit tightly into the bowl. According to Mandler, this is what contributes to the difficulty of learning verbs: relational concepts are packaged differently across languages.
Choi studies these phenomena in Chapter 7, examining the concepts related to containment and support events of prelinguistic infants in English and Korean cultures. Before infants learn the appropriate labels of their language, they distinguish between tight-fit and loose-fit containment events, regardless of their culture. On the other hand, they do not categorically distinguish between tight-fit and loose-fit support events (described with put on in English). Only older Korean infants distinguish the latter. Choi hypothesized that it is due to the large variability between the different kinds of support events.
Based on similar findings, Casasola, Bhagwat and Ferguson, in Chapter 6, argue that some relational meanings can be easily formed regardless of linguistic abilities. However, language experience contributes to the development of the meanings of support and tight fit.
Another instance in which languages package concepts differently is motion verbs. Verbs describing motion events in which a figure moves along a path can either refer to the manner of motion, or to the path traversed. Motion verbs in English predominantly refer to the manner of motion, whereas in Spanish, motion verbs commonly refer to the path of the motion. In order to learn the appropriate meaning of motion verbs, infants first need to break motion events into their component parts. In Chapter 5, Pulverman, Hirsh-Pasek, Golinkoff, Pruden and Salkind report that children as young as 7-months-old detected changes in manner and path in events in which an animated starfish was moving relative to a stationary ball. In addition, 10- to 12-month-olds categorized events based on path.
In Chapter 8, Buresh, Woodward and Brune point out another conceptual challenge to be considered for verb learning. Verbs refer not only to the actions of an agent but also to the desires and goals underlying those actions. Thus, learning the meaning of verbs would depend on whether infants can reason about an agent's goals and desires. In this chapter they review findings that suggest that in their first year of life infants can reason about the intentions of others. However, they point out that infants produce such verbs only a year later. Therefore, they argue that infants may be able to interpret the intentions of others in order to reason about events but might not be able to extract exemplars from these events so that they can be mapped onto verbs. In Chapter 10, Poulin-Dubois and Forbes find that two-year-olds do not readily extend novel verbs to new events that differ in the actor performing the action. They argue that children learn verbs related to actors' intentions later in development.
In Chapter 11, Behrend and Scofield suggest that there are reciprocal relations between children's understanding of intentions and verb learning. Not only do children need to understand others' intentions in order to learn verbs, but also hearing an actor label his actions helps clarify his intentions. When an actor labels his actions it gives a cue that the action is intended rather than accidental.
Another conceptual challenge for verb learning is action segmentation. Just as words are not separated by silences in the speech stream, actions are not separated by clear pauses. In order to infer the correct concept from an event, infants need to know which actions belong in ‘one structural gestalt’. In Chapter 9, Loucks and Baldwin examine this and find that 10-month-olds' segmentations of events are similar to those of adults.
Even though infants in their first year of life demonstrate rich conceptual knowledge, verbs are still hard to learn. The following chapters offer insights as to why this is so by comparing verb learning with noun learning. In Chapter 21, Gentner presents her well-known theory that nouns are easier to learn than verbs. She argues that the first nouns children learn are easy to learn because they refer to objects and animate beings, which are concrete and highly imageable. On the other hand, even some imageable verbs, for example, those describing motion events, are harder to learn because languages vary in how they refer to these events.
Another reason why children might learn nouns before verbs is that the meaning of some verbs depends not only on the action but also on the object involved in the action. This is the main thesis proposed by Kersten, Smith and Yoshida in Chapter 19. They find that children's success in learning and extending verbs depends on their familiarity with the object. They also point out, however, that verbs differ in the extent to which they are dependent upon properties of objects.
In Chapter 18, Tardif proposes that the noun/verb asymmetry in acquisition is influenced by language-specific characteristics. For example, child-directed speech in English contains more nouns with specific meanings as compared with Mandarin. The other pattern is true for verbs. In Mandarin verbs in child-directed speech refer to more specific meanings as compared with English. Therefore, Mandarin-speaking children might find it easier to learn verbs as compared with English-speaking children.
In addition to linguistic differences, cultural variables also influence noun and verb learning. In Chapter 20, Lavin, Hall and Waxman propose that Easterners might show an advantage in verb learning because they show a preference for relational categories, while Westerners show a preference for taxonomic categories. To test this hypothesis, they asked adult subjects to guess the words an American mother was saying to her child in videotaped interactions. They found that Canadian college students guessed more nouns than verbs relative to Japanese students. Yet both groups were more likely to guess correctly for nouns than verbs.
So, how hard is it to learn verbs? In Chapter 12, Childers and Tomasello found that 2·5-year-olds needed more exposures to a novel verb than a novel noun in order to produce it. The children had no problem performing the action that the verb labeled, rather, they had difficulty performing the action and labeling it at the same time. However, verb production improved when presented in longer, more naturalistic sentences. Yet younger children, two-year-olds, did better with nouns in comprehension and production, even with longer sentences.
Although Childers and Tomasello succeeded in teaching novel verbs to two-year-olds, not all studies succeed in doing so. In Chapter 14, Maguire, Hirsh-Pasek and Golinkoff found that 18-, 24- and even 30-month-olds failed to learn novel verbs. This is probably due to differences in the task. Here, the children were asked to extend a novel verb to a situation involving a new actor. Even seeing the action performed by four different people did not help them extend it to a new person. They explain these findings by proposing that all words, nouns and verbs, fall on a single continuum determined by shape, concreteness, imageabillity and individuation. Where a word falls in this continuum determines how difficult it is to learn.
Another possible interpretation of these findings is that it is not the case that children always learn nouns before verbs, but that the properties of their input language, English, facilitate noun learning. In order to test this hypothesis, in Chapter 17, Imai, Haryu, Okada, Lianjing and Shigematsu compare the performance of English-, Mandarin- and Japanese-speaking three- and five-year-olds in learning novel nouns and verbs. In all three language groups the three-year-olds associated the noun with a novel object, and the five-year-olds but not the three-year-olds associated the verb with an action. However, it was not the case that the three-year-olds thought that the verb referred to the novel object; rather, they thought that it referred to the interaction between the action and the object.
In Chapter 13, Naigles and Hoff demonstrate that it is not the case that very young children cannot learn verbs at all: 18-month-olds successfully extended some familiar verbs to a new situation in their experiment. Yet they point out that it is hard to estimate which and how many verbs children know in general with the current methods of input sampling. They inspected interactions of mothers with their children and found that the mothers did not produce a representative sample of the verbs children use early in development. For example, mothers never said clap, jump, roll or wave during nine hours of interaction with their children, performing tasks that were designed to elicit those words.
What are the useful cues children might use to learn verbs? The syntactic bootstrapping theory suggests that sentence structure can inform verb learning. In Chapter 15, Fisher and Song propose a mechanism for early syntactic bootstrapping. In particular, they describe how children identify the subject of the sentence, telling them who does what to whom. They suggest that children map word order onto a very abstract notion of prominent participant in an event.
In Chapter 16, Lidz provides further evidence for syntactic bootstrapping, finding that children used verb transitivity to infer a causative meaning. This was the case regardless of whether transitivity was the most reliable cue for causation in the language. Not only did English-speaking children use the number of arguments as a guide to verb meaning, but also children speaking Kannada, a language that has a causative morpheme which is a more reliable cue to causal meaning. This suggests that children are guided by sentence structure to infer the meaning of a verb.
These twenty-one chapters elucidate on the various factors that contribute to the challenge of verb learning, taking into account different perspectives. What can we conclude, in spite of these challenges, about children's verb knowledge? On the one hand, findings in this book demonstrate that very young children succeed in learning verbs: two-year-olds can comprehend and produce newly learned novel verbs, and 18-month-olds can extend some familiar verbs to a novel experimental setting. This is consistent with findings published since the publication of the book. These findings suggest that children as young as 21-months-old have abstract syntactic knowledge that allows them to interpret novel verbs (Gertner, Fisher & Eisengart, Reference Gertner, Fisher and Eisengart2006; Conwell & Demuth, Reference Conwell and Demuth2007; Fernandes, Marcus, DiNubila & Vouloumanos, Reference Fernandes, Marcus, DiNubila and Vouloumanos2006) and that by five years children not only know some verbs, but know enough to expect the kind of arguments that can occur with them (Snedeker & Truesewell, Reference Snedeker and Trueswell2004). On the other hand, other findings in the book suggest that verb learning is hard. For example, under different settings five-year-olds but not three-year-olds succeed in extending a novel verb to a new situation. This exemplifies how complex the verb learning process is. One reason for this discrepancy might be that understanding a verb's meaning depends on determining what perspective the speaker took. In some settings this might be easier to do than in others, depending, for example, on the sentences the verb was presented in, and on how salient the goal of the action was. These questions can be further examined in future studies.
The book is a very useful reference on the topic of verb learning, clearly describing different perspectives that need to be considered. It would be an excellent resource for a graduate class on language acquisition. The chapters are written in an accessible way, presenting the theoretical background and data in a clear manner. It is thought-provoking and provides much motivation for continuing to do research in the field.
