Language Acquisition Lab, Department of Linguistics
University of Massachusetts Amherst
Projects for 2013-2014
1. Show me the lion next to the hippo under the gorilla. Is it the hippo or the lion under the gorilla?
Jon Nelson, from our Spanish linguistics department, studied the interpretation of sentences with multiple prepositional phrases like the one above. Such sentences can typically be understood as either coordinate, in which the lion is both [next to the hippo] and [under the gorilla], or as recursive, in which the lion is [next to [the hippo under the gorilla]], that is, the hippo is under the gorilla, not the lion. Jon asked children, ages 5 through 8, as well as adults to arrange animals on a tablet to match his sentences.
Do children place the lion or the hippo under the gorilla? How does their understanding change as they get older? And how do their interpretations compare to adult interpretations of the same structures? In the next phase of his study, Jon will do the same experiment with Spanish children and adults.
In his work this spring, Jon found that across all ages, children and adults preferred recursion (with the hippo under the gorilla) to coordination (with the lion under the gorilla), although the difference in the preference across ages was substantial. While the youngest children chose a recursive interpretation only 45% of the time, that number steadily increased to 70% among the adults. Likewise, while the five-year-olds preferred coordination 32% of the time, this steadily decreased to less than 20% with the adults. Furthermore, the coordinate answers were not evenly spread across the study group, but instead were largely produced by a small number of participants across all groups, suggesting that there may be something akin to a subconscious ‘Aha!’ moment when a speaker begins to prefer recursion.
Recursion has been proposed as the core feature of human language and one of the characteristics of universal grammar that ties all languages together. However, much work remains to be done to investigate how recursive ability is acquired by children and adults learning specific languages.
2. The popstar kicked himself (at the wax museum)
We are used to thinking that if “John kicked himself” he must have kicked his own body, not someone else’s. Now suppose John kicked a statue portraying him. Did he kick “himself”?
Valentina, a visiting researcher from Italy, is looking at this syntactic puzzle from an acquisition perspective. Do four- and five-year-olds understand this special meaning of reflexives like “himself”? Do their grammars start out with a very narrow definition of “himself” which includes only the subject in the flesh? Or do they start with a broad definition that could include every representation of the subject that is present in the context?
To answer these questions, Valentina designed stories around Disney characters and their statues, in which she changed the set of alternatives, characters and events. For example, in one type of scenario, Cinderella falls in the mud with her statue and wipes the statue off with a towel. In other scenarios, Cinderella has two statues, one of herself and one of another person. Both need cleaning--but Cinderella herself doesn’t, and she cleans only the statue portraying her. Children then were asked two types of questions: “did Cinderella dry herself off?” or “did Cinderella dry off?”
Four- and five-year olds often gave “himself” a broad meaning and accepted “Cinderella dried herself off” as a truthful description of Cinderella drying her statue. Interestingly, they rejected “Cinderella dried off,” which is similar in meaning but slightly different in its syntax, for the same scenario. This showed us that the children’s grammar could make distinctions based on the form of the reflexive. Much like adults, 5-year-olds were more prone than 4-year-olds to say “no, she dried off her statue!” or even “she dried off her other self!” They accepted a broad meaning of “herself” only when the real princess was not among the alternatives.
If we start as children with a broad interpretation, how do we, as adults, narrow down the meaning of “himself”? Does it relate to our capacity to project sets of alternatives? And why can some reflexives stretch their meaning in context while others do not? This experiment is only a first step but it contains some important new insights to take these questions further.
Each and every again (and again and again)
Three studies this year looked at how children learn the meaning and uses of each and every. Do the two words have the same meaning? Can they be used interchangeably?
3. Did Suzy carry “every” ball or “each” ball?
Jeremy, Magda and Tom explored whether children ages at ages 5 to 7 distinguished “each” and “every” in stories about events. In their stories, they contrasted events that happened in a sequential order or simultaneously. After the story was told, two puppet helpers, Phil and Matt, said what they recalled from the story, and the children had to decide which puppet was correct.
For example, in one of the sequential stories, Suzie and her friend want three balls. First Suzie carries a white ball over to her friend. Then she went back and took the yellow ball and carried it over to her friend, and finally Suzie went back and took the green ball and carried it over to her friend. In the simultaneous version of the story, Suzie would carry the three balls at once. After each story, the puppets said what they thought happened,
Phil said: Suzie carried each ball over to Mary.
Matt said: Suzie carried every ball over to Mary.
Each child heard just one version of the scenarios, but heard both descriptions from the puppets. For the story where the balls are carried one by one, adults indicated that both each and every were appropriate. They chose one or the other, or said both were correct, and children did exactly the same.
Adult and child responses were different, though, when the stories described actions happening all at once.
For simultaneous events, adults only found every appropriate, but the children’s preference was the opposite: both 5- and 7-year-olds preferred each 75% of the time. The children seem to have mastered how each and every are the same, but were still working out how they differ.
4. “Every bear is tickling a turtle,” “Each bear is tickling a turtle,” and “All the bears are tickling a turtle”: How many bears? How many turtles?
Margreet, our Dutch visitor, also worked with “each” and “every” and contrasted them with “all.” She asked American English-speaking children and adults and Dutch-speaking children and adults to choose between a picture of three bears tickling one (the same) turtle and another with 3 bears each tickling a different turtle for “all”, “each” and “every.”
With sheep riding elephants, crocodiles pushing tigers and other fanciful events, Margreet first confirmed with adults where the Dutch and English languages differ. The Dutch words “elk(e)” and “ieder(e)” do not distinguish “each” and “every”, but should be closer to “each” based on their properties. Dutch-speaking adults and children both overwhelmingly preferred the picture with 3 bears tickling different turtles for “elk(e)” and “ieder(e)”. They even preferred that same picture for “alle”, although to a somewhat lesser extent.
American English-speaking adults were similar to Dutch adults for “each” and “all.” They preferred the picture with 3 bears tickling different turtles for “each,” and they preferred the 3 bears and 1 turtle situation for “all”. But what about “every”? Adults accepted either an “each” or an “all” interpretation for “every,” but in Margreet’s experiment they showed that they preferred the “each” interpretation (3 bears tickling different turtles) although less so than the Dutch adults did. American English-speaking children through age 9 did not seem to have made up their minds on “every”: they picked either picture just as often as the other. They do not seem to have formed a preference for treating it like “each” and are more likely than adults to treat it like “all”.
So, even though English and Dutch resemble each other on many points of grammar and semantics, there are still many differences between these seemingly similar words. Many questions remain, and the next step for Margreet is to look at English-Dutch bilingual children’s preferences for “every.”
5. Is every girl riding a bike? Does it matter if there are more bikes than girls?
Amanda and Jeremy are trying to find out why children treat the universal quantifier every differently than adults do. As an example, if you ask, “Is every girl riding a bicycle?” where there is a picture of 3 girls riding a bicycle and an extra (empty) bicycle, adults and teens say “yes,” but many children will say “no, not that bicycle.”
Children, ages 4 to 8, seem to know that every quantifies over individuals (like “girl”), but as in the example above, they do not always restrict the “domain” of every so that it applies only to individuals in the same noun phrase (“girl”). The domain seems to include “bike” as well.
One hypothesis to explain this non-adult interpretation is that children let the visual or discourse context—not the syntax--determine every’s domain of quantification so that it includes “bicycles.” Another explanation that has been offered for children’s responses is that they treat every as a quantifier over events, as if the question was “Is every event an event of a girl riding a bicycle?”
To decide between these hypotheses, Amanda and Jeremy presented 34 children, ages 4 to 8 years, with a series of pictures and sentences, and asked them to judge whether the sentence was true or false based the picture. For example, they asked them “Is every cowboy riding a horse?” about a picture of three cowboys each riding a horse and an extra horse with no cowboy riding it. Each child did the task in two sessions. In one session, the “every”-sentence was preceded by a lead-in question that focused their attention on the correct domain of quantification, like “How many cowboys are there?” In the other session, no lead-in was presented.
They found that children’s responses were more often adult-like when the test sentence was preceded by a lead-in. They propose that if children treat “every” as a quantifier over events, the discourse context (which highlights individuals) should not cause them to change their analysis of the event and give adult-like responses. They believe the results support the hypothesis that children have access to the adult interpretation of every as a quantifier over individuals, but rely on context rather than the syntax of its sentence to restrict its domain.
6. His and hers –or is that his and his?
Terue Nakato-Miyashita of Japan, who spent a year at UMass in 2011, visited us again this fall to go to the schools where she had worked before. This time, she investigated if English-learning children consistently use “his” only for a boy (or a man), not for a girl (or a woman). For example, when we say “Mary is brushing his teeth,” we normally understand that Mary is brushing some boy’s teeth, not her own. She found that English-learning children, even around the age of seven, do not necessarily use “his” only for a boy: they tend to (or sometimes almost completely) ignore “gender information” when their preference for some interpretation is enhanced by visual or contextual information.
Terue showed pictures where a boy and a girl were doing some action. For example, one picture showed a girl is pinching her own nose and a boy is pinching her nose. Then she asked “Is Green (= the boy) pinching his nose?” 40-to-50% of the children said “yes” to the question, which suggests that they did not necessarily rely on gender information of “his” in the question. Interestingly, this tendency got much stronger in the following situation: a girl is pinching her own ear and a boy is also pinching his own ear. When asked “Is Blue (=the boy) pinching his ear?”, almost 90% of the children said “no.” This again shows that they ignored gender information of “his”.
Her interpretation of the results is that it takes longer than we previously thought for children to learn the use of “his” or “her” in a completely adult-like manner. This might not be so mysterious, however, because adult-English has some exceptional use of pronouns, where a male pronoun can refer to a female. For example, if we say “Every man voted for himself, and Becky did, too.” Here the second sentence can be understood as “Becky voted for Becky, not every man.” Children might use the same strategy adults do, but not necessarily in the same contexts as adults.