Do you agree with this phrase? “Some elephants have trunks.”
If you’re a pragmatic thinker, you might disagree. It’s not just some elephants that have trunks; all elephants have trunks. If you’re a logical thinker, you might agree. All elephants have trunks; therefore, it follows that a portion, or some, of the elephants have trunks.
Now, what if we paired that initial phrase with a context cue? What if we asked, “Do any elephants have trunks?’ Or “Do all elephants have trunks?” Would that change how you think about the word some?
For many people, that extra context does have an impact. That’s according to new research by Teaching Assistant Professor Glenn Starr, who studies experimental pragmatics, psycholinguistics, and second-language acquisition in UWM’s Linguistics Department. He just published a paper in the journal Quaderni di Linguistica e Studi Orientali: Working Papers in Linguistics and Oriental Studies that examines how people understand the word some based on context.
For his study, Starr asked a group of native English speakers to read two-sentence conversations, paired with a picture that showed five identical objects, like five apples. The first sentence contained a context cue. For example, it might read, “Did you eat any of the apples?” Or, “Did you eat all of the apples?” The corresponding answer was, “I ate some of the apples.” Based on the picture, the participants had to tell how natural or appropriate the response was.
“In English, some usually means not all, so if the picture showed that all five apples were eaten, saying ‘I ate some’ might sound a little strange, especially if the question used the word all,” Starr explained. “However, with any in the question instead, the salience of the not all meaning is diminished, and it is now possible to interpret some logically as compatible with all.”
While the paper revealed some interesting insights into how people use the word, what the research really reveals, Starr said, is how our cognitive control strategies impact how we interpret language.
Here are three things you should know about Starr’s latest research.
1. “Some” is a complicated word.
Some has two meanings. It most often refers to a certain amount of a whole, but we also use some to mean an entire group. For example, in the sentence “I have some friends who enjoy bowling,” some specifies the whole group that likes knocking down pins. And, some exists on a scale with other words that refer to amounts. Few, many, most, and all are scale-mates of some, Starr said.
“When you use some, usually in conversation, people will derive a ‘some, but not all’ inference,” Starr added. “That’s above and beyond its strictly literal interpretation. You have to take the literal interpretation and then derive extra pragmatic meaning out of that.”
So, the question then becomes, do people automatically infer the pragmatic ‘some, but not all,’ when they hear the word some? Or do they usually default to the semantic, literal interpretation of some to mean a whole group?
“This is important because we want to know as human beings whether we derive certain types of information first and then, through extra effort, do we arrive at an alternative interpretation? How do we derive inferences in human communication?” Starr said. “This little word some is an intriguing and simple way to test this mechanism.”
2. In his research, Starr found that the people interpreted some based on context.
The research showed that people decided that the some response was less natural when the preceding question used the word all rather than the word any. That meant that the context of the scenario influenced how people interpreted the word some in the answer sentence.
Starr also paired the language portion of the test with a pattern-matching test that measured how well participants could keep track of context. Surprisingly, said Starr, older adults who scored lower on the pattern-matching test were more sensitive to the context words.
Why? It comes down to cognitive control strategies, which are ways of maintaining context. Younger people generally have a proactive cognitive control strategy, Starr said: “When they take in contextual information, they hold it in working memory, and they can dispel that information to solve task-relevant goals.”
But as we age, our ability to hold information in our working memory is diminished. For example, a young person might remind themselves throughout the day that they need to buy groceries after work; an older person might forget until they see a shopping list which jogs their memory. This is a reactive control strategy. “They activate context maintenance cues only when they come upon some kind of confusing or ambiguous information, and then they think, ‘What am I supposed to be doing? Oh, yeah!’” Starr said.
The experiment results suggest that, instead of constantly keeping track of context, those older people tended to wait until something seemed confusing and then carefully examined the context cues. Interestingly, the pattern-matching test could capture this strategy, and the resulting behavior was reflected in older participants’ tendency to interpret some in the different ways mentioned above.
3. The results have implications for how we communicate with each other, especially as we age.
Proactive and reactive cognitive strategies are not set in stone; plenty of young people have reactive cognitive strategies, and vice versa. But overall, Starr said, the results of his experiment show that we should be mindful of how we communicate as we age.
“Older people may need to be reminded, or maybe they need to have more routine awareness of these cues in order to interpret inferences,” he said.
The impact of this research goes beyond aging. Starr said his results have implications for how people learn a second language. Certain second language-learners often have strong attentional control ability, which also translates into increased context sensitivity. Thus, this line of research could help language teachers tailor curriculum to their students. The results could also have implications for how people with autism – generally a population that struggles with interpreting pragmatic context – infer meaning in conversations.
And, said Starr, these results are a reminder of how we use quantifying language to describe our surroundings. “It’s such an integral part of how we carve up our realities,” he said. “Whenever I hear the word some, I think, is that an existential some, or a pragmatic some, and what cues … led (the speaker) to that interpretation?”
It gives us all some things to think about.
By Sarah Vickery, College of Letters & Science