Abstract:

Acquiring vocabulary when learning a language is a critical part of reading comprehension and verbal expression. Traditional modes of vocabulary acquisition often rely on rote memorization techniques such as the use of flashcards. These highly-criticized techniques dissent with the widely-supported Incidental Vocabulary Learning Hypothesis. The hypothesis supports gradual, repeated exposure to words through extensive reading as the most productive mode of vocabulary acquisition. This method is ideal for non-specific vocabulary acquisition goals. However, learners are often presented with targeted vocabulary learning goals, such as is the case with spelling tests and standardized exams such as the SAT/GRE. There is currently no known scalable method for learners to choose a book to read that will best meet targeted learning goals.

This paper describes a prototype of a virtual tool called the Vocabulary Acquisition Simulation Tool (VAST) that takes English literature as input and estimates the relative comprehension that a person will attain of each unique word by reading that text. VAST can also work in reverse by using a list of vocabulary words as an input to query several different books and outputting a choice of book that will best improve comprehension of the vocabulary list.

VAST relies on a simulation technique known as Nodal Formulation that can be used to model a variety of complex engineering, science, and socio-economic systems. Applying this technique enables the process of language learning via reading to be modeled as a network of nodes, where each node is a unique word. These nodes are connected by a set of governing equations that relate different parameters which are hypothesized by literature from educational psychology to impact vocabulary acquisition. Parameters considered include the orthographic complexity of a word, the frequency with which a word appears in the text, etymological relationships between words, and an individual subject’s propensity to learn and retain words. A variety of computational techniques are applied to increase the speed of the tool.

The accuracy of the tool is evaluated using a spelling test to perform a controlled before-and-after study, comparing the post-reading spelling test results to our predicted results. VAST has potential to:
(i) assist educators in choosing required readings,
(ii) replace manual methods of assessing educational interventions on vocabulary acquisition, such as the Peabody Picture Vocabulary Test and
(iii) empower learners – from English Second Language learners to students studying for standardized tests – to have more sovereignty over their particular vocabulary acquisition goals.

Keywords:

Language, learning, vocabulary, reading, technology.