Researcher in Profile: Marisa Casillas on Children’s Interpretation and Understanding of “Almost” and “Sort of”

By Chia-wa Yeh, Head Teacher and Research Coordinator

“Cow!” a 3-year-old exclaimed, pointing to the picture of a bison. Her parent said, “It’s sort of a cow,” and went on to explain that it’s really a bison. How might the child interpret the information modified by the phrase “sort of”?

“Sort of” is what linguists call a hedge phrase. In the example of the bison, the parent went on to point out the similarities and differences between a cow and a bison. By using a hedge phrase, parents offer information about specific features associated with a new word. Instead of saying “No, it’s not a cow,” parents often take an indirect approach to lessen the impact of an otherwise negative statement while at the same time providing new information.

Hedges are words that signal vagueness or fuzziness in categorization (Lakoff 1973, Kay 1997, Lasersohn 1999). Adults often use them to extend children’s vocabulary or to explain the function of unknown objects.

Are young children sensitive to adults’ use of such hedge phrases? If so, at what age are they picking up on these hedges as being useful? Do they make inferences based on the information when they do nocasillas those words?

Marisa (Middy) Casillas, a fourth-year Stanford graduate student in linguistics, her advisor professor Eve Clark, PhD, and Patrícia Amaral, PhD, an assistant professor at the University of North Carolina at Chapel Hill, investigated children’s acquisition of hedges last year and are currently conducting a follow-up study on this subject. Amaral had conducted a study looking at children’s understanding of the adverb “almost” at Bing Nursery School while a post-doctoral fellow at Stanford. Casillas and Amaral then decided to collaborate on a study to look further into the domain.

Casillas grew up in Walnut Creek in the East Bay. Part of Casillas’s interest in conducting research with children stems from being the oldest of four children and having the experience of taking care of her brother, who is 11 years younger. Casillas graduated from the University of California, Los Angeles, majoring in linguistics and psychology. She is athletic and enjoys sports. Her favorite is rugby, which she says could easily have been her third major in college, as she played it all four years. Casillas also enjoys baking, drawing, quilting and reading.

In this study, Casillas and Amaral focused on children’s understanding of “almost” and “sort of.” Casillas played a guessing game with 3- to 5-year-olds in which she placed a picture of an object in a box and then had a bear puppet peek inside the box and give children clues about the pictured object. In random order, the bear gave one of four descriptions as clues. The four descriptions were: “It’s almost a ______” (hedged frame); “It’s sort of a ______” (hedged frame); “It’s a ______” (unhedged frame); “In here there’s a ______” (unhedged frame).

Casillas then showed children four pictures and asked them to make a guess as to which of these pictures resembled the one inside the box. The children indicated their guess by pointing to one of the pictures.

The sets of pictures the researchers used examined the children’s abilities to understand hedging in two different conditions: “prototypicality” and “completeness.” Here’s how that played out when “butterfly” was the image hidden in the box. To test the children’s understanding of hedges given a prototypical choice, the four images shown were a butterfly—the prototypical object—a moth, a bumblebee and a praying mantis. The moth is considered a “competitor” as it shares many properties with the butterfly but is not a butterfly. The bumblebee and the praying mantis, both insects like the butterfly but bearing no close resemblance to it, are “distractors.” To study their understanding of hedges when the hedge related to “completeness,” the researchers replaced the image of the moth with a picture of a cocoon with a partially emerging butterfly.

Casillas recorded all the descriptions to ensure children heard the utterances in exactly the same length and tone from the bear puppet—which entailed an intricate set-up. The researcher placed a speaker inside the bear with a cord attached to an iPod she hid out of the child’s view.

The researchers hypothesized that if children understood the meaning of “sort of” and “almost,” they would be more likely to choose the non-prototypical competitor (e.g., moth and cocoon with a partially emerging butterfly) as opposed to the prototypical item (e.g., butterfly) and distractors (e.g., bumblebee and praying mantis) when they heard a hedged frame (e.g., “It’s sort of/almost a butterfly”) in the clue. If children heard an unhedged frame (e.g., “It’s a butterfly” or “In here there’s a butterfly”), they should be more likely to choose the prototypical item. The researchers also predicted that there would be a developmental effect of the interpretation of hedges—that younger children would choose prototypical objects (e.g., butterfly) when they heard a hedged frame (e.g., “It’s sort of a butterfly”) as opposed to their older counterparts.

The results of this study showed that 5-year-olds responded in almost the same way as adults when they heard hedged phrases. Adults chose prototypical competitors (e.g., moth) 100 percent of the time; 5-year-olds did 75 percent of the time, whereas 3- and 4-year-olds did so less often. However, after reviewing the children’s response delays (i.e., hesitation time) and comments, the researchers found that 4-year-olds were at an intermediate stage in which they took nearly as much time when responding to hedged frames as 5-year-olds, even if they eventually chose the prototypical object. The researchers believe that this suggests that 4-year-olds are taking the hedge into account to some extent even though they have not yet fully grasped its meaning.

Casillas and Amaral also found the unexpected result that 3-year-olds appeared to clue into the length of the descriptions and treat long unhedged phrases as if they were hedges. The long unhedged frame (e.g., “In here there’s a butterfly”) was purposefully constructed to match the length of the hedged word/phrases to avoid potential reactions based on the length of the sentences alone since the other unhedged frame (e.g., “It’s a butterfly”) is two syllables shorter than the hedged ones. The researchers suspect that rather than cueing into the meaning of the lexical terms such as “almost” and “sort of,” 3-year-olds paid more attention to the length of the description.

Casillas and Amaral wrote a paper on this study, which will be included in the Proceedings of the Berkeley Linguistics Society. They have received enthusiastic responses at conferences and are conducting a follow-up study at Bing to look deeper into children’s acquisition of the two hedges.

In the follow-up study, the researchers use made-up objects created out of paper shapes and label them with made-up names (e.g., “hup,” “shan”) to eliminate the need for children to have prior knowledge of the target objects in order to respond correctly (e.g., to know that butterflies emerge from cocoons). Using these made-up objects enables the researchers to establish what is prototypical and what is complete to the children. The rest of the study is the same as the first one.

The researchers suspect that the results will be similar to the first study. Such results will provide additional support for the conclusion that children’s responses were based on their knowledge of and attention to the hedge phrases.

“Research at Bing is a pleasure,” Casillas said, and added that she has enjoyed interacting with the children. The children, who found her talking bear quite intriguing, have enjoyed her research too.