Ivy J. Livingston
As Christopher Francese notes in the introduction to the Dickinson Core Vocabularies <http://dcc.dickinson.edu/vocab/core-vocabulary>, teachers of Latin and Greek have long recognized that novice students will benefit greatly from first learning the most frequently occurring words in the language. More recently, research on second language acquisition shown that vocabulary, rather than grammar, is the most difficult aspect of reading for language learners (Day and Bamford 1998:78). But how should one go about learning those most common words?
In his summary of the research, Folse (2004:39-45) concludes that both intentional learning (e.g. from definition lists or flashcards) and incidental learning (e.g. from repeated encounters in the course of reading) are effective, with intentional learning being most efficient for initial familiarization and incidental learning being necessary to develop a sense for how words are used, to learn collocations, and to prevent loss. But in order to learn new language items from context and/or to deepen knowledge of vocabulary and grammar already encountered, learners must be able to read relatively quickly and easily, without frequent recourse to dictionaries. It has been estimated that at least 90% of the words in a reading should be already known to the student.
In developing Hedera: A Personalized Language Learning Environment, the project team wanted to be able to calculate the readability of a Latin or ancient Greek text from any source with a high degree of precision for any individual learner, taking into account that students do not typically learn, for example, ab (place from which) and ab (personal agent) at the same time. To achieve this, it was necessary to develop a system that would enable users to go beyond digital lemmatization that associates forms with the headwords of a given dictionary. To give a more specific example: a student learning Latin from the Cambridge Latin Course might have learned nearly 1000 words by the end of the third volume, including iratus (defined as an adjective “angry”) but not the verb irascor. If iratus is lemmatized as a form of irascor, then we cannot capture the fact that our hypothetical student would actually understand the sentence Abiit iratus.
In this session, I will show how Hedera (1) enables teachers to customize the lemmatization of texts and (2) allows learners to track not only what words they know (even if the word is not considered a headword in the standard dictionaries), but also how well they know each word. I will then demonstrate how these two strands work together to determine with greater accuracy whether a text will be readable enough to support incidental learning and the reinforcement of known words. Hedera also facilitates intentional learning by listing unknown words and exporting them for learning on paper or with digital flashcards, giving learners the option to gain a basic familiarity with any unknown words before beginning the text. Finally, I will show how Hedera’s reading environment also tailors support to the individual learner.