StudyDB is a database interface designed to create student knowledge bases by means of tracking study habits. This tracking information can help students sort through a database to find the study material that they need to learn. As students study, the database keeps track of time and context which can be later used for review and for generating drills and exercises customized to each student’s knowledge base.

 Memory in the Information Age: New Tools for Second Language Acquisition
(published in the ADFL Bulletin, Winter 2003)

How many of us remember our bank account numbers? Telephone numbers of collegues? Driving directions? Thanks to technology, we no longer need to recall such trivial information, we can look it up on our laptops, our PDA’s, our cell phones. As global positioning system (GPS) technology is added to our mobile computing devices, we will no longer need to remember where we are, or where our loved ones are (Specter) Not certain what you are looking at? Steven Feiner describes the augmented reality (AR) that can be achieved through a head mounted device that annotates what you see by combining GPS information with databases to superimpose labels on the objects in your perceptual field. New technologies may well make it unnecessary to learn foreign languages! Already there are handheld devices that use speech recognition to translate words and common phrases (Breenan).

Does ready access to such information atrophize our ability to remember? David Shenk argues that more and more people today are overloaded with information, a veritable data smog that surrounds us whose symptoms include attention deficit disorder, high blood pressure, declining visual acuity, impaired judgment, overconfidence (36-38). As well, he argues that information overload can indeed lead to memory loss. He attributes this to the loss of diverse situational cues, to getting so much of our information from the same place, our computer monitor which becomes the primary context for so much of the information we receive (48).

Still, if the above describes ways in which technology eliminates the need or reduces our ability or need to remember certain types of information, there are many ways in which technology can be an aid to memory. One promising approach that has emerged in recent years is tracking technology. Tracking keeps a record of how a user interacts with information presented to them on the screen. It has been used primarily to analyze user habits for pedagogical or marketing purposes and/or evaluating the effectiveness of a computer interface. When this tracking information is stored in web accessible databases, it becomes possible to analysis the information interface interaction patterns of a group of users. Many online assessment tools employ tracking technology to record a student’s answers, the date they took a test and the amount of time they take to complete the test.

StudyDB
What will be described here is a database project currently under development at Middlebury College which is called StudyDB. StudyDB is set of relational databases with an interface that allows students to create custom study lists from items they find in the database. It is essentially like flashcards, only much more flexible. Like flash cards, students can drill themselves, separating and sorting the individual items into different lists. The database keeps track of the number of times a student sees a given item, the last time they saw the item and of all the times they drilled themselves on a given item, the percentage of those times which they were able to recall or recognize it. A student can then use this information to sort their custom study lists in much the same way they would sort flashcards.

Such a methodology, that of direct instruction of vocabulary, attracts considerable criticism from those favoring the communicative approach, with its emphasis on communicating messages and creating meaningful contexts in which to acquire language; and those who argue that vocabulary is best acquired through reading. Yet recent studies show that some direct instruction is indeed helpful (Ellis, 1994; Hulstijn, 2001; Zimmerman, 1997) and this is especially true for beginners. Laufe has proposed that for English, at least 3000 word families (words and their most usual derivations and inflections) need to be known for the basic reading comprehension out of which new vocabulary may be acquired. Groot has noted too that for intermediate and advanced stages of language acquisition, students often need to learn a large number of words in a short period of time, words that cannot easily be acquired incidentally, as they do not occur often enough in foreign language learning material (61).

What is being suggested here is that we can use the same technology that is available for retrieving information as a means for aiding the retention of that information. The search interface of StudyDB is designed for both the presentation of lists of vocabulary items and/or phrases organized by units for direct study and the presentation of glosses for words a student looks up while reading L2 texts. The interface can be configured to present words/phrases/glosses in either L1 or L2 or in stages beginning with L2 context sentences and followed by L2 glosses, before finally providing L1 glosses. The tracking technology remembers every item a student identifies as known or unknown in direct study or look ups while reading. With additional tracking information such as retrieval date, frequency and context, students can sort their study lists by date last seen, number of times seen or knowledge (ratio of times seen to times recognized, recalled or identified as known).

The StudyDB interface has three different modes of presentation. The first is the simply drill mode. Here a student can specify what to display, either L1 or L2 text. As well, a student can chose to display contextual information. This could be a image, a sentence showing the vocabulary item used in context, or any other information that can be linked to the item. If the student chooses to hide additional contextual information, it can be retrieved by a mouse click if necessary. Audio playback too is always available via a mouse click. Then the student has the option of adding any item to their known or unknown list. This is the mode of presentation that is most similar to traditional flashcards and might constitute the first mode of study for new vocabulary/phrases or would be the principal mode for glossing words while reading.

The second mode of study is recognition. StudyDB can create various types of multiple choice quizzes based on search and display options in the database interface making it possible for students to practice recognizing the meaning of an item as one of a set of possible choices. These multiple choices are randomly selected from the database according to the search criteria enabling a form of contextual hinting in vocabulary acquisition. Thus a student could chose this mode to study new items from a given lesson. If the lesson is organized around words and phrases of a particular semantic field such as food or family or imperatives, then the recognition drill will require students to identify the correct match for a given food word/phrase from other words or phrases that are also related to food. If, following Foucault, we regard words and things as defined not by some essence, but by recognizing what they are not, what distinguishes them from all other things like them, then this may be a powerful and efficient way to gain L2 mastery of a given semantic field.

A student can also review their own study lists (lists of words that have been identified as known or unknown) in this recognition mode. Doing so means they would need to recognize items from among other items they have already seen. Such a drill can sharpen ones existing vocabulary, clarifying what has already been learned. Recall too that a given student’s study list can be sorted according to when they last saw a given item, the number of time they have seen an item and how well they know the item. Thus a student can chose to focus on new or old vocabulary from the perspective of time, frequency and knowledge.

The last mode of study is recall. Here instead of selecting the correct choice from multiple choices, a student is required to type the correct L2 translation of given L1 item or vice versa (much improvement can be made here, since the matches need to be exact according to the database, allowing no variation in possible translations). This mode can function as a cloze exercise, particularly if a student chooses to show an item in L1 with a context sentence in L2 (the L2 item targeted in this display configuration is replaced by a blank).

Discussion
Is all of this merely "drill and kill"? There are mind numbing elements to this systematic display of vocabulary items. If we do decide to relegate certain uses of StudyDB to such a categorization, we ought to consider at least that it is very efficient, customizable, extensible "drill and kill", far more flexible than learning vocabulary via bilingual lists or traditional flashcards. For less commonly taught languages, there seems to be a great need for such tools. The web abounds with various flashcard programs, particularly for Chinese and Japanese. What makes StudyDB unique at this point is that it is connected to a web-based database. This means that an instructor can view the study lists of all students and monitor their progress. In additional to being useful for drilling, StudyDB is also a bilingual dictionary that can be used to look up unknown words encountered in L2 reading material.

We have no idea how effective StudyDB will be for students. We plan to beta test it this summer. We are not aware of any other software with exactly the same features, though we have modeled some of its features on existing software on the market. In particular, we have incorporated features from the CAVACO (Computer-Assisted VOCabulary Acquisition) program that is described by Groot. We have developed another flashcard-like program called FlashTrack that has been used by the Chinese, Japanese and Arabic schools at Middlebury College for the last couple of years. Testimonials from students suggest that program was relatively useful. What distinguishes StudyDB from this earlier program is that it is database driven and can maintain records of students performance across sessions, so that students could develop extensive study lists over time and call up specific information about how they have interacted with the database. The various display, sorting and study modes make it a surprisingly sophisticated tool whose effectiveness will depend as much on how skillfully its options are employed. It is conceivable that the interface itself could analyze students study habits and suggest study methods. Commercial packages that offer similar functionality do not give as much control over the display and/or sorting of items. They do this perhaps because they have worked out an optimal methodology or more likely to keep the interface simple and easy to use. While there is much to be said for simplicity, I think it is ultimately more important to develop interfaces that allow for the most control possible over the available content.

This interface is a primitive example of the real power of technology to process information in general. We are all familiar with knowledge bases; these are essentially databases containing the collective knowledge of many individuals who have some expertise in a given topic or the knowledge that is required in a given course of study. What we have here is the beginnings of what we might describe as a student knowledge base. It is a database of the what a student has learned. Tracking technology of this sort could replace traditional testing as we know it. Instead of requiring students to all take a particular test within a particular time period, students could simply be required to log so many hours of time using one of these interfaces. This is the notion of continuous testing. Since everything is stored in databases that are web accessible, instructors could monitor the study lists and quiz results of the class as a whole or each student with the computer generating relevant statistical information.

A new technology that has gained popularity in recent years, particularly in the private sector, is collaborative filtering (Gladwell). This is a technique for predicting preferences based on comparing individual choses with those of a population. For example, all of your purchases at amazon.com are entered into their databases and your purchasing profile is compared with others. From that comparison, the database then tries predict what other products that might interest you since those products were purchased by individuals whose profile was similar to yours. It is as uncanny process that seems to be intelligent. In fact it is simply pattern matching across a population

The same technique could be applied to learning. As students use tools such as StudyDB, their individual study statistics can be compared with the study statistics of other students. From this comparison, the technology could generate Krashen’s comprehensible input +1, finding just those words or phrases that extend a student’s knowledge by means of combining the familiar with the new. Thus students who know the words for apples and oranges should be taught the word for fruit which could naturally lead into vegetables and foods in general and eating and cooking and hunger and desire and so on.

Works Cited
Breeden II, John. "Handheld translator speaks four tongues." Government Computer News. November 13, 2001.

Ellis, N. Consciousness in second language learning: Psychological perspectives on the role of conscious processes in vocabulary acquisition. In J. Hulstijn & R. Schmidt (Eds.), AILA Review, 11 (1994), 37-56.

Feiner, Steven K. "Augmented Reality: A New Way of Seeing". Scientific American, April 2, 2002. Gladwell, Malcolm. "The Science of the Sleeper". The New Yorker, October 4, 1999.

Groot, P.J.M. “Computer assisted second language acquisition”, Language Learning and Technology Vol. 4, no.1, (2000): 60-81.

Hulstijn, J. H.. Intentional and incidental second-language vocabulary learning: A reappraisal of elaboration, rehearsal and automaticity. In P. Robinson (Ed.), Cognition and Second Language Instruction. (2001): 1-21 Cambridge University Press.

Krashen, S. We acquire vocabulary and spelling by reading: Additional evidence for the input hypothesis. The Modern Language Journal, 73, (1989): 440-464.

Laufer, B. What percentage of text-lexis is essential for comprehension? In C. Lauren and M. Nordman (Eds.) Special language: From humans thinking to thinking machines. Clevedon: Multilingual Matters, 1988.

Shenk, David. Data Smog: Surviving the Information Glut. New York: HarperEdge, 1997.

Specter, Michael. "No Place to Hide". The New Yorker, November 27, 2000. Zimmerman, C. Do reading and interactive vocabulary instruction make a difference? An empirical study. TESOL Quarterly, 31 (1), (1997): 121-140.

Zahar, R. Cobb, T. & Spada, N. Acquiring vocabulary through reading: Effects of frequency and contextual richness. Canadian Modern Language Review, 15(4), (2001): 541- 572.