I Did Some of My Homework. I Did All of It, Too.

Both Stiller and Barner present research about the phenomenon in which, compared to children, adults better understand meaning implied by scalar quantifiers, stating the historical understanding in the literature that pragmatic inference grows over time. Both papers acknowledge that children’s difficulty with scalar implicature is a failure to come up with relevant lexical alternatives to the quantifiers used.

However, the studies recognize that children’s difficulties are not easily attributable to what was suggested in previous literature such as limitations of memory or issues with demands of the task at hand; indeed, children in both studies were able to make correct inferences in context-dependent questions where they were able to construct ad-hoc scales.

What becomes interesting to me is how, using these insights, one might alter early childhood education to address this gap in knowledge of lexical alternatives. Might teachers stress the scalar nature of quantitative words that are not inherently numerical, highlighting the implied mutual exclusivity of “some” and “all” for instance – that when someone eats “some of the cake” it conventionally excludes the idea that they ate “all of the cake?” 

Personally, I still struggle knowing how to scale more nuanced words like few, some, and several, which leads to another question of how one might teach a computer to quantify and order such terms when there is no context of jelly beans or balloons as visible representations. Since it's not explicitly stated, how might the computer infer semantic meaning and become capable of recognizing implications?