LEE FORESTER
Hope College
Abstract:
A basic understanding of general research on human memory can help guide CALL design to new innovations. Though this statement seems self-evident, most CALL software does not exploit research in psychology to the fullest extent possible. In addition, many CALL software reviews do not consider issues of cognition and memory in their analyses. This article offers a brief overview of what is generally accepted about how human memory works as it applies to CALL. Such a basis can help both CALL designers and users develop a more sophisticated understanding of what CALL software can and should do in supporting learning. A number of interactions from various CALL products will be discussed in the light of the research summarized here.
KEYWORDS
CALL Design, Human Memory, Human Perception
BACKGROUND
Before discussing what is known about human memory, it is useful to establish two foundational concepts that will inform the rest of this article. The first is the fundamental distinction between experiential cognition and reflective cognition. This distinction is known by many different names and is key to any understanding of human memory and learning. Experiential cognition represents skilled activities that can be conducted without pausing to think. Athletes, musicians, police officers, and pilots all use experiential cognition when they respond to situations automatically. Experiential cognition is relatively effortless, though it requires extensive practice to achieve adequate levels of proficiency in a field.
Reflective cognition represents thinking which requires effort and time. It is used for solving complex problems and addressing new or anomalous situations and benefits from mental tools such as books, pencils, print, language, and mathematics. Much of what is normally conceived as education is considered part of reflective cognition. Both experiential and reflective cognition are often mixed as in video games which require reflective problem-solving as well as experiential game-play. Language learning can be seen as moving from primarily reflective cognition, where learners struggle with creating a sentence, to more automatic experiential cognition when one is fluent in a language and can respond without effortful thought.
99
The second foundational concept is the understanding of the computer as an information medium, like print, video, audio, film, CD, and television. The computer is commonly understood metaphorically as a tool, and, while there is some validity to this notion, it is incomplete and at times misleading. Each information medium, including computers, has its own strengths and weaknesses, its own "truth." Print can provide a strong support to reflective cognition, and fields such as literature and philosophy would be very different without it. Television, on the other hand, is well suited to experiential cognition since it represents actions and processes through time and requires no effort, but it is notoriously poor at fostering deep reflective thinking. Such authors as Marshall McLuhan, Walter Ong, and Neil Postman have written extensively on the differences between print culture and electronic culture. These same concepts need to be applied not just to television but also to computers.
The popular term "computer multimedia" is thus something of a misnomer. The word multimedia itself indicates a combination of media such as text, images, audio, and video. With it, computers are understood as a combination of other existing media. However, computers are not simply a combination of older media (now including radio, telephone, and television). They represent a medium of their own with unique strengths, weaknesses, and cultural settings. Currently most development on computers seems to be repurposing materials from other media into computer format. While this kind of development is usually a step forward, until the medium of the computer is understood in its own right, progress will be limited. We do not yet know all the aspects of computers as medium, and indeed we probably will not recognize them until the next great shift in information media. What is required now is a conscious, persistent questioning of assumptions about how computers work, realizing that we often transfer concepts from previous media without validating them in the computer environment.
HUMAN MEMORY AND LEARNING
Human memory has a physiological basis in the human brain, and learning involves chemical and physical changes in the brain. While the process is not well understood (and leaving aside discussion of the mind-brain question) learning should not be viewed as some sort of mysterious spiritual event. Memory and learning seem to be the result of more efficient neural circuits that arise from rich and extensive use. Learning and memory are sensitive to the physical state of the learner and can be influenced by fatigue, biological cycles, stress level, time of day, chemicals and drugs, and a host of other factors.
Two basic kinds of learning observed in humans and virtually all other organisms are habituation and sensitization. Habituation is a form of learning in which an organism responds less strongly to a repeated stimulus, effectively "tuning it out." Sensitization is the opposite: a strong stimulus leads to a stronger response to similar stimuli. Thus, becoming startled by a loud sound may make a person "jumpier," and a fairly quiet sound may evoke an overly strong response.
100
Habituation is very useful since it helps organisms focus only on new and important stimuli. For example, habituation allows us to drive while thinking about something else. With computer interfaces, habituation is useful because users can navigate through software without having to concentrate on the computer, focusing instead on content. However, a form of habituation can also occur when learning software has the same format, the same activities, and the same layout through various units. This template approach is often advocated because it saves development time and costs, but, because of habituation, it can lead to decreased attention which, in turn, inhibits the formation of long-term memories (i.e., learning). CALL software should encourage habituation with the interface but not with the content so that learners respond to the content of the software with interest and attention to the greatest degree possible.
KINDS OF MEMORY
There are four basic kinds of memory: sensory memory, short-term memory, working memory, and long-term memory. When external events are perceived, these perceptions are first stored briefly in sensory memory. We pay attention to some of this input, and it moves to short-term memory which can last several seconds but rapidly deteriorates. During this time, some of the information may enter working memory, in which input can be processed and compared to old knowledge. If information is processed more fully, it may finally proceed to long-term memory which handles all memories that are longer than several seconds. Naturally, any trace can be strong or weak and disappear or become inaccessible in long-term memory; not all memories are equally strong.
Sensory Memory
Sensory memory is used to create a whole sensory picture. It is usually fairly brief and allows us to combine images and sounds. For visual stimuli, iconic memory stores an exact representation of what is seen but lasts only about 0.2 seconds, after which the memory trace disappears. Iconic memory can also include an afterimage which can last several seconds under ideal conditions (a bright stimulus preceded and followed by dark). Playing with sparklers at night is a great way to see an afterimage—rather than a single point of light, we perceive lines of light. Film and video also take advantage of iconic memory. With film, the screen is only illuminated half the time, and, with television, only half of the image is ever displayed at one time. Our iconic memory stores the previous image, and thus we perceive both film and television as continuous images.
Echoic memory is used as a temporary store for auditory input. It lasts a bit longer than iconic memory, and words or other interpreted signals are usually stored for 2-4 seconds. Echoic memory helps us understand sentences as we store a few seconds worth of speech. Neither iconic memory nor echoic memory
101
requires attention; thus, the common occurrence of saying "What?" and simultaneously 'hearing' in echoic memory what was just said.
As a side note, care must be taken when designing repetition tasks in CALL because improperly designed tasks may require only sensory memory. Sensory memory is effortless, does not require attention, and disappears very quickly without any learning occurring. An auditory trace can last a few seconds in echoic memory so activities that require a learner to hear a word and then immediately link it to the word in print may rely purely on sensory memory, that is, with virtually no deep processing. Likewise learners can repeat what they hear into a microphone and then completely forget it as the trace disappears from the auditory store without even making it into short-term memory.
Neither visual nor auditory perception are monolithic wholes. Vision is processed in separate areas in the brain according to color, motion, form, and depth, and each of these visual subsystems behaves differently. The color system has fairly low resolution, which can lead to problems in software when color text is placed on an equiluminescent color background. Even though the colors may differ, if their luminescence is the same, we must rely on the color visual system to see the difference; because this system has low resolution, the text is often completely illegible, a problem that can be seen on some web sites. In addition, human language and nonlinguistic sounds are processed differently by the brain.
Short-term Memory
Short-term memory (STM) is used to combine sensory inputs and to focus on them for a short time. STM dissipates quite rapidly, lasting only a few seconds, unless effort is expended to rehearse or encode the information. The following are a number of key concepts associated with STM:
1. Digit Span (the number of items that can be held in STM)
While digit span varies from individual to individual, it is generally 7 plus or minus 2 (so 5-9 items). The term comes from psychology experiments testing how many random digits can be recalled in sequence with a 50% chance of success.
2. Chunking (the organization of items into familiar, manageable units)
Chunking can occur automatically or through effortful encoding. Chunking helps extend the digit span: One can recall 7 single-digit numbers, 7 words (composed of several letters), or perhaps 7 phrases. There is a hierarchical nature to chunking in which more information can be remembered (stored) in STM by grouping it into wholes or chunks.
3. Rehearsal (the process of repeating information to keep it in STM)
Through rehearsal one can keep a phone number or other items in STM longer than a few seconds. Rehearsal takes effort and can be interrupted by other demands on STM.
102
4. Proactive inhibition (PI) (the interference of one item with other items)
When trying to store several items in STM and to encode them in long-term memory, each item can interfere with others. The more similar items are to each other, the more they can interfere with learning successive items. PI can be acoustically based; thus, it is more difficult to recall a list of words if their sounds are similar. PI can also be triggered by semantic similarity; thus, lists of similar items (e.g., lists of fruit) are generally more difficult to memorize because the semantically related items interfere with one another.
5. Recency Effect
The most recent item or two in a list are usually recalled much better than other items.
It is essential to note that these are features of short-term memory, and may or may not apply to long-term memory.
Working Memory
Working memory provides a somewhat larger view of STM which is currently in vogue in cognitive psychology. It focuses on the actual processes in STM, how it is broken down, and how it connects to previous knowledge. There are three major parts of working memory: The visual spatial sketchpad, the phonological loop, and the central executive.
1. Visual Spatial Sketchpad
As the name suggests, this stores visual and spatial information (which are not the same thing). Multiple tasks that require visual and/or spatial processing conflict with one another. Thus, it can be difficult to envision a complicated chess match in your mind while driving. However, it is not overly difficult to think about a chess match and talk about it simultaneously since speech usually does not place any demands on the visual spatial sketchpad.
2. Phonological Loop
This part of working memory seems to have the purpose of assisting in language acquisition. It is sensitive to speech versus nonspeech sounds and is one reason why it is extremely difficult to attend to two conversations simultaneously; it is usually not difficult to listen to someone in the presence of nonverbal sounds (e.g., music playing in the background at the same time as a conversation).
3. Central Executive
The central executive coordinates the visual spatial sketchpad and phonological loop as well as other aspects of working memory to form a whole gestalt.
One implication of the working memory framework is that the visual spatial sketchpad and phonological loop can function simultaneously without overburdening STM. When creating language-learning software, it is a good idea to
103
keep both of these channels in constant cooperation, for example, using audio to support imagery/text and vice versa, and avoiding overloading either the visual spatial sketchpad or phonological loop with competing information of the same type.
Long-term Memory
Long-term memory (LTM) involves relatively permanent memory traces that are stored beyond STM. These traces can last from a few minutes to an entire lifetime, depending on how deeply the traces are processed. Two kinds of memories stored in LTM are explicit memories and implicit memories. Explicit memories are things that can be consciously recalled, while implicit memories occur without conscious knowledge. Explicit memories include facts, general knowledge, and personal experiences, all of which can be recounted. Implicit memories include motor and cognitive skills as well as conditioned responses and learning. In many cases, experiments with amnesic patients show that they still have strong implicit memories, even though they have no explicit recall of ever having learned them.
Psychology research has determined a number of features of explicit memory and learning that have direct bearing on language learning and CALL.
1. Rate of Learning
Other factors being equal, the rate of learning of new material is fairly linear. The more time spent, the more is learned.
2. Distributed Practice
Distributing learning in smaller 'chunks' spread out over time is more efficient than learning all at once. This concept meshes with the conventional wisdom of language teachers and has been borne out by psychology research.
3. Role of Attention
Attention is crucial to learning. Without attention and semantic processing, long-term learning will not occur; getting and keeping the learner's attention is a must for effective CALL.
4. Role of Motivation
According to most research, motivation does not seem to play a direct role in learning. It naturally plays an indirect one in that it assists in attention and effortful processing, but motivation itself does not correlate with learning. As long as learners are actively processing information, their motivation does not have a significant impact on their measurable learning.
5. Role of Repetition
Repetition with attention is good, but pure repetition by itself does not increase LTM. It will often take frequent repetitions to learn new information, but, again, frequent repetition only helps with attention.
104
6. Role of Successful Recall
Successful recall strengthens the neural pathways thereby assisting in moving information into LTM. Successful recall is the proper role of repetition. The more something is successfully recalled, the better it is remembered; successful repetition can have a positive effect on learning.
7. Role of Predictability
Predictable environments also assist in learning. For language learning, predictable environments include context and background schemata, both of which have received prominent attention in the literature on L2 acquisition and learning.
8. Role of Arousal
Learning occurs best when people are somewhat aroused (not asleep!), but not too agitated. In addition, humans tend to be very interested in and aroused by the unusual, the bizarre, the surprising. Arousal is a key to the television news broadcasts in which one strange human interest story after another vies for our attention. The usual is not recalled very well; the unusual is often recalled extremely well. This is also the case with shocking or deeply emotional experiences, which can often be vividly and accurately recalled many years afterwards.
9. Level of Processing
The more senses that are involved in processing an item and the more semantic processing that is involved, the better the chances that a piece of information will be successfully encoded in LTM. If one simply reads a word, recall is not very good. Reading the word and hearing it (multiple senses) increases the likelihood of successful recall. Reading the word, hearing it, and organizing it according to its meaning yields an even better chance of successful recall and learning.
Compared to explicit learning, implicit learning plays a much lesser role in CALL. However, it should not be completely ignored. Two important features of implicit learning that can be successfully used in CALL are the following:
1. Priming
Priming is the unconscious retrieval of an association. If one hears or sees a word, it is more readily recognized later, even if the priming stimulus cannot be recalled at all. Thus, for example, if vocabulary items are briefly introduced before a listening comprehension task, they will show priming effects such that the learner will hear the items more readily, even if the items cannot be recalled later and even if the effect is not consciously noticed.
2. Evaluative conditioning
People tend to prefer the familiar to the unfamiliar. This preference applies to various stimuli and is not a conscious process. In other words, when subjects
105
are presented with two items, one of which has been previously presented, they will usually prefer the one previously presented, even if they cannot recall having seen it before. This type of learning can be exploited by exposing learners to extensive cultural information about the societies of the L2 being studied. Even with minimal explicit learning (e.g., facts, figures, and names), a great deal of implicit evaluative conditioning can occur, making the learner more positively inclined towards native speakers of the L2 and perhaps more prone to continue their language study. Unfortunately, few CALL products aim at this sort of exposure to authentic language and culture, preferring instead inauthentic audio and visuals in order to cut costs.
RECALL
Just because something is in LTM does not mean that it can be recalled successfully. Recall can be assisted in numerous ways:
1. Cues
Recognition is much easier than recall. Often, if we have forgotten someone's name, we can recall it when we are given the first letter of the name or we can select the correct name when given a list of five names. However, without these cues recall might not be successful. The difference between recall with and without cues is important to recognize in designing CALL interactions. CALL does a good job using cues such as audio files or sounds to facilitate recall, but it has not been as successful at strengthening free recall without cues.
2. Context
Information can be more easily recalled in the context in which it was learned. Consequently, items should be learned in multiple contexts to strengthen learning. The use of multiple contexts takes more processing but leads to deeper learning. In addition, care should be taken when evaluating learning in a context different from the one in which it was learned. For learning materials, the importance of context would speak for the explicit and consistent use of multiple media (e.g., print, computer, and video) rather than reliance on one medium alone.
3. Emotional States
Emotional states can also prime recall. Depression, anxiety, and stress block recall. This is in accord with the perceived role of the affective filter in communicative language learning. Pleasing, interesting environments are more conducive to recall than boring, stressful ones, an idea that must be borne in mind when creating CALL software.
106
SOFTWARE ANALYSIS
Based on some of the psychological factors outlined above, we will now turn to several brief analyses of a number of old and new CALL products, focusing in the main on mechanisms used for glossing and vocabulary acquisition. We are not taking a theoretical stance on the value of glossing in L2 learning, but since glossing is a standard item in CALL software, various approaches will be examined to see to what extent they exploit features of human memory.
The Neue Horizonte NOW! CD-ROM was produced by Transparent Language to accompany Houghton and Mifflin's textbook Neue Horizonte. Figure 1 shows a typical view of the software.
Figure 1
Neue Horizonte NOW!
0x01 graphic
The program contains only text and audio; there are no images. This particular approach treats multimedia as multiple open books. The advantage is that users can access the text, translation, notes, and grammatical features within one interface, clearly a time-saving device if users need this information. However, the open books metaphor does not take advantage of what is known about human memory and perception. Scanning between the various items takes time and energy; there is no highlighting or visual directing of attention to salient help texts. The power of imagery is completely sacrificed.
In terms of memory, the program treats the computer essentially as print rather
107
than as its own medium. It strongly favors reflective cognition and does not exploit both reflective and experiential modes. The drab, repetitive interface leads to habituation. The audio (available through clicking on any text) does not complement cognition but only offers pronunciation. If grammatical or other notes were available in audio, that availability would allow simultaneous visual-spatial and phonological loop processing in working memory. Chunking is not exploited, nor is there any rehearsal or review. Recall is not checked at this stage, though vocabulary games elsewhere in the program address recall to some extent. Priming and evaluative conditioning are not considered.
Fortunately, this program represents more where CALL has been rather than the direction it is going, but it should be clear that this sort of interface and approach do not take full advantage of what the computer can do for language learning. It has speeded access to helpful information, but that information is still steeped in print culture and print thinking.
The Practice Makes Perfect German application was produced by The Learning Company and has been unavailable for some time. The application shows some definite improvement, particularly in the use of images and a friendlier layout (see Figure 2).
Figure 2
Practice Makes Perfect German
0x01 graphic
The textboxes in the lower left are in much better juxtaposition, and the text is visually isolated from interference to allow easier scanning. Five images on
108
screen fits the low end of 7 ± 2 for chunking, a better choice than the typical four as seen in products such as The Rosetta Stone. This software uses voice recognition, although it does not perform exceptionally well.
A few things can be improved. The omnipresent textbox is easy to program but breaks up the visual design. As on newer television commercials, text does better when it is blended into the visual space, not pasted on top of it in a box. The images also bleed into each other perceptually and should be differentiated from one another, not simply juxtaposed. Audio is available for pronunciation and is used to elicit recognition, but recall itself is not practiced. Levels of processing are not exploited, and the inauthentic imagery does not support implicit learning. Culturally appropriate ambient sound is also lacking. These two elements could provide tremendous exposure to the target culture, strengthening implicit learning through evaluative conditioning and priming of cultural artifacts
The Let's Talk German program by Syracuse Language has an approach similar to that of Practice Makes Perfect German. It employs voice recognition, though again with mixed results. In Figure 3, a word in the target language is presented via audio, and users must identify it by clicking on the appropriate image. Variations on this theme can be found in numerous programs. This type of interactivity can be good, but it has some pitfalls as well.
Figure 3
Let's Talk German
0x01 graphic
109
The program tests recognition, not recall, and is strongly subject to process-of-elimination strategies. The task is not to understand this word but rather to click on the correct box. If the word heiß 'hot' is spoken, perhaps users understand it, perhaps not. If users know kalt 'cold' and glücklich 'happy,' only two possibilities remain, allowing for the substantial possibility of guessing. Half the time, users can get the correct answer without any higher level linguistic processing of the problem. Indeed, the item is probably not even rehearsed in working memory for any length of time, just enough of it to identify whether it matches the pictures or not. It could even be processed as "I know it/I don't know it," which would suffice to solve this problem.
Using imagery works well for concrete objects but not so well for abstract concepts. The image in the upper left hand corner can be identified as 'cold' if it had been primed that way, but it could also mean 'freezing,' 'two people,' or something else. The image in the lower left hand corner probably means 'chest,' but other options are available as well. Without priming, how can users really know what the image in the lower right means? Woman? Smile? Shy? Happy? Just because we can use images does not mean we should use images.
Finally, to foster cultural learning and positive evaluative conditioning, authentic imagery should have been used in the program.
The Learn to Speak series, created by The Learning Company, has progressed to version 8.0 and is one of the most well established CALL applications available. In version 6.0 of the German program pictured here, there are 48 different units, each dealing with a particular pragmatic situation geared toward travelers. Each unit is structured in the same way, progressing from vocabulary through various exercises and activities. This repetition is easy to program but can lead to habituation of content. A general goal in CALL should be a varied visual presentation that keeps interest without losing habituation of the interface.
The focus on vocabulary takes many forms, which is good. Words are chosen from alphabetical lists (an artifact of print) but then appear in a separate box with a gloss beneath it and an associated image. In addition, users can immediately link to the context in which the word occurs. Linking to a context is distinct advantage because it changes the cognitive task, requires more in-depth processing, primes the word for listening comprehension, and increases the chances of successful recall. It also strengthens connections to cues and context which assists in later recall. The program's interface is rather busy. The vocabulary notes do not need to stay on screen. They can be made available, if desired, but simply clutter the display if they are always present. The image of the woman speaking is small and is set against a blue screen with no visual interest or sense of authenticity (see Figure 4).
110
Figure 4
Learn to Speak, 6.0
0x01 graphic
Having all the lesson controls available does not mean that they all have to be displayed at all times. A great deal of space is yielded here to the interface and keeping all information available at all times. The constant presence of the lesson controls interferes with focus and attention. A better approach would be to remove all the boxes and move most of the elements off screen, making them available through mouse-based roll overs, buttons, or some other mechanism. The increased screen simplicity would greatly enhance the visual appeal and keep users more cognitively engaged.
The Einblicke series of CD-ROMs is based on the video series of the same name. These videos were repurposed for Fokus Deutsch, and these CD-ROMs could be used to supplement that series. The glossing method here is one that calls up a dictionary. Users find the word, either by typing or scrolling, click 'OK,' and then the gloss appears with the word in a context along with its translation. Having a context directly available as well as the audio is a definite advantage, though the interaction is somewhat cumbersome. Multiple clicks and two boxes are required to see the gloss. The method of bringing up separate boxes is efficient for programming but breaks the flow and removes the gloss from its original context. To minimize the disruption of the learning flow and extensive visual scanning, glossing should usually be an integrated part of the program, not an additional window that breaks users' concentration and makes them close the window when finished (see Figure 5).
111
Figure 5
Einblicke
0x01 graphic
The initial screen of Learn to Speak French, 8.0 has been much improved from earlier versions (see Figure 6).
Figure 6
Learn to Speak French, 8.0
0x01 graphic
112
The interface is more attractive and modern, showing a better understanding of visual communication. Use of a modern sans serif font and clean imagery helps greatly, and the transparent geometric background imagery does not distract the user. The organization is also improved, with pronunciation offered up front and the basic course being separated from the more pragmatic travel course. This organization in which content is modified according to users' needs is something that computers can easily handle and is much more difficult in print. Internet material is incorporated directly into the interface—a positive development. This version contained a copy of Fodor's Guide to France, a cassette tape, and a print workbook. The use of multiple media supports different levels of processing, priming, cueing, and recall.
Improvements have also been made in glossing and word look up. There are three tabs from which to choose. The first tab is expressions (not words), grouped according to topic (not alphabetically) with the option of translation and direct drilling with Automatic Speech Recognition (ASR) and playback (see Figure 7).
Figure 7
Expressions Tab
0x01 graphic
Notes are also given, and an image displayed with the appropriate section in the 'dialog.' Users can also access individual words through the words tab and expanded tab—grouped according to semantic topic (see Figures 8 and 9).
113
Figure 8
Words Tab
0x01 graphic
Figure 9
Expanded Tab
0x01 graphic
114
The interface is certainly more attractive than in other products, though it would still be better to avoid splitting the screen into set regions for set functions that simply rewrite the changing content. Nevertheless, this approach has the advantage of easier programming and a consistent interface, so perhaps it is an acceptable tradeoff here.
Overall there have been some very positive changes in version 8.0 that begin to take better advantage of unique aspects of the computer medium. At the core, we still have lists and print, as indicated by the text boxes. We will soon need to move beyond this to a more integrated and organic approach. To do so will take a different view of design and also require programmers to redesign the basic approach. Just because text boxes are easy to implement does not mean that they should always be used.
A few basic design considerations would include the following:
1. The screen display should show only what users need at that moment. A dynamic interface (one that changes according to user inputs) decreases habituation and increases attention and depth of cognitive processing.
2. Elements on screen should usually be separated by white space to reduce visual processing time and effort. In the figures above, scanning from the top of the vocabulary box to the notes takes some effort because users have to pass over numerous intervening items, even though those items are helpfully separated by color and shape .
3. Chunking should be used. Presenting users with long lists of vocabulary does not usually help. In the above example, chunking is implemented in the number of items in each topic, but ideally it should also be used visually (i.e., no more than 7 ± 2 on screen at a time).
4. Image and text need to be closely associated. They should ideally be available in the same visual saccade/scan and not separated across the screen if at all possible.
5. Textboxes are not always helpful. When working at the word level, we can learn much from television practices in which producers know how to use text in a synchronous and dynamic, multimedia environment.
Moving from standalone software to courseware, the Nuevos Destinos CD-ROM is certainly one of the best to come out. It follows up on and supports the Destinos and Nuevos Destinos video series for beginning to intermediate Spanish learners. The overall approach is to engage learners in solving problems in Spanish through role playing, one of the more promising approaches to education that multimedia can offer.
In Figure 10, glossing of underlined items is available via a mouse click. The English gloss appears in a box right above the word in question, and users know which words can be glossed by the very light red underlining. This approach is very integrated and does not interfere with users' experience or significantly impact the visual field. The context stays visible; the gloss fits into the context
115
well, and the gloss appears at the point of attention allowing users to easily see it with only one saccade.
Figure 10
Nuevos Destinos CD-ROM
0x01 graphic
The size of both phrases is small enough to require only one eye fixation for reading. One minor improvement could be made by changing background shape/color of the gloss box to aesthetically match its background through transparency or some other device so all the original text is still available. Yet, this design would require more programming, and, with limited budgets, it is not a high priority. Overall, this technique is psychologically an excellent approach to glossing.
The Portes ouvertes application (Holt, Rinehart and Winston) is a first-year college French program integrating a textbook and multimedia CD-ROMs in one package. The sample screen in Figure 11 shows one approach to studying vocabulary that works well for concrete objects. Items belonging to the concept of desk are arrayed on an image of a desk. Clicking on an item highlights the item and brings up the French audio of the word. When the 'T' button is depressed, the French text appears unobtrusively below the desk. This technique works well because users first scan the item and then hear the French, using both the visual spatial sketchpad and phonological loop. If users do not understand the French, the task is NOT to link the object and the French word but rather to segment the French word. In that case, clicking on the object produces
116
the audio, and, since the object is already primed and stored in STM, the visual spatial sketchpad can be used to dissect the audio via the written text.
Figure 11
Portes ouvertes CD-ROM
0x01 graphic
Reading the text of single words (within one eye fixation) while hearing the audio uses both major channels together and is an excellent learning technique, especially because it is learner-instigated and thereby correlates well with a high degree of attention. Because each semantic set has a different image associated with it, the imagery changes (defeating that dreaded habituation) and serves a primary function. This interaction demonstrates good utilization of psychological principles
Short video clips are subtitled below the clip when the 'T' button is depressed. The subtitles are of the correct length, containing one or at most two eye fixations, and are placed directly below the image, making it easy to scan to the text without intervening obstructions. Until wider use is made of keywords and text on the actual screen, this approach is superior to the standard method of having a textbox relatively far from the actual video window (see Figure 12).
117
Figure 12
Video with Subtitles
0x01 graphic
In the preparation activity, users can click on images in study mode to hear the people introduce themselves (see Figure 13).
Figure 13
Preparation Activity
0x01 graphic
118
In test mode, the audio segments are played in random order, and users have to remember to which person the audio belongs. The idea of this interaction is good, but eight alternatives are very high on the 7 ± 2 scale. There is also a great deal of proactive interference; since the introductions are fairly similar, it is easy to confuse them. This activity places a very high demand on short term memory because without a great deal of knowledge of French, individual introductions probably cannot be chunked; STM can get filled with just one or two introductions. Learners will have to rely on other strategies (e.g., identification of gender, accent, or background sound). These are valid skills, but perhaps not the ones in need of reinforcement. To focus on that learning strategy, one could play an audio segment and have students guess who said it. This approach would engage higher order cognitive skills and bring in cultural perspectives/stereotyping, although it would not be an exercise in linguistic recall. Essentially, the preparation interaction places too great a demand on STM.
In the expansion activity, users play the video, listening to a description of hobbies, and then click on each hobby that is mentioned (see Figure 14).
Figure 14
Expansion Activity
0x01 graphic
The context is given (we know that it is about hobbies) and looking at the images gives users time to build associations and prime the French words (that users know) beforehand. Then, the task of discriminating the words from the spoken audio is not so difficult. This exercise is not mediated by English, which
119
is positive, not because English mediation is necessarily bad, but a mix of modalities is cognitively appropriate, including activities without English. The expansion activity is a psychologically nice interaction.
The following screens are from a prototype program for beginning German entitled Auf Geht's! (Live Oak Multimedia). In the screen in Figure 15, learners step through the gloss of the Einstein quote, which is accomplished in chunks.
Figure 15
Sample Glossing Procedure in Auf Geht's
0x01 graphic
The background text fades in order to enable easy scanning from German to English. The gloss color (dark red) is different enough to draw attention but does not hurt the eyes or cause visual fatigue. The sentence is divided into coherent units to implicitly demonstrate German syntax. The objective here is that users will look at each of these chunks in isolation and perhaps develop some questions (e.g., about word meaning and word order). This method of glossing needs to be tested with learners to see if it is successful, but it at least represents a computer-based approach rather than a print-based approach because it uses the dimension of real time presentation and sequence in the same visual space, something that print cannot do.
The screen shown in Figure 16 is designed to prime vocabulary for listening comprehension.
120
Figure 16
Priming Vocabulary
0x01 graphic
Users click on a word, hear the audio while the print is still visible, and, while the elements are still in visual and audio STM (afterimage and echoic memory), the printed gloss appears in the same spot (see Figure 17).
Figure 17
Glossing Primed Vocabulary
0x01 graphic
121
Ideally, students should rehear what they hear subvocally when the gloss appears, taking full advantage of multiple stimuli in STM. The German text also pops back to increase binding to the English text that was read and perhaps rehearsed in STM as well. This experimental approach attempts to utilize features of memory but has not yet been validated experimentally; it is unclear whether the advantage of forced rehearsal is offset by other negative factors, such as short duration of the gloss.
For the interaction shown in Figures 18 and 19, clicking on the word replaces it briefly with the English gloss.
Figure 18
Timed Glossing Procedure: A
0x01 graphic
Figure 19
Timed Glossing Procedure: B
0x01 graphic
122
In this mock up, users can modify the length of time the gloss stays on the screen by typing in the numbered textbox at the lower left hand corner (1 = 1/60 of a second). Most users can recognize the word down to a value of 10 (1/6 second). A value of 6 (0.1 seconds) is the shortest time most people need to recognize the first letter of a word. If users try a setting of 5 or less, the original afterimage blocks perception. A speed of 15 or so should work fine.
Experimenting with time of presentation is interesting because near subliminal glossing may turn out to be effective in some contexts. A short presentation forces more active rehearsal in STM but may also not allow enough time for processing. A modified version of this technique plays an English audio gloss rather than showing the textual gloss. Some teachers hesitate to endorse English language audio in such an environment, but psychologically it plays to the strength of working memory because the German text (visual-spatial sketchpad) does not interfere with the English audio (phonological loop). Experiments with these and other interactions are planned to see what really happens with CALL learners in these new situations.
Another type of glossing is proposed here, based on the work of Nagata (1999). The idea is to use semantic processing to assist learning in glossing. When users click on the word to be glossed, two choices become available (see Figure 20).
Figure 20
Multiple Choice Glossing Procedure
0x01 graphic
123
Users must then choose the correct gloss based on the context. When designing such glossing pairs, the meaning must be somewhat close or at least feasible to encourage deeper processing. Presenting two wildly different possibilities would limit thought because the answer would be obvious. There seems to be a lot of promise with this method in using glossing to facilitate word learning, not just text comprehension; it results directly from memory and language learning research and keeps users more active.
CONCLUSION
In many ways, our understanding of the principles of CALL are still in formation. It is hoped that this short outline of salient factors from research in human memory and their application to specific CALL applications will be of assistance both in designing effective CALL software and in evaluating current CALL projects and products.
SELECTED BIBLIOGRAPHY
Baddeley, A. (1999). Essentials of human memory. Hove, England UK: Psychology Press/Taylor & Francis (UK).
Baddeley, A., Gathercole, S., & Papagno, C. (1998). The phonological loop as a language learning device. Psychological Review, 105 (1), 158-173.
Borchardt, F. L. (1997). Theory, assessment, praxis: On the body and soul of CALL. In K. A. Murphy-Judy (Ed.), Nexus: The convergence of language teaching and research using technology (pp. 13-27). Durham, NC: CALICO.
Brown, C. (1993). Factors affecting the acquisition of vocabulary: Frequency and saliency of words. In T. Huckin, M. Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 263-286). Norwood, NJ: Ablex Publishing Corporation.
Cheung, H. (1996). Nonword span as a unique predictor of second-language vocabulary learning. Developmental Psychology, 32 (5), 867-873.
Chun, D. M., & Plass, J. L. (1996). Effects of multimedia annotations on vocabulary acquisition. The Modern Language Journal, 80, 183-198.
Coady, J., & Huckin, T. (1997). Second language vocabulary acquisition: A rationale for pedagogy. Cambridge, U.K., New York: Cambridge University Press.
Cohen, R. L. (1988). Metamemory for words and enacted instructions: Predicting which items will be recalled. Memory & Cognition, 16 (5), 452-460.
Ellis, N., & Beaton, A. (1993). Factors affecting the learning of foreign language vocabulary: Imagery keyword mediators and phonological short-term memory. Quarterly Journal of Experimental Psychology, 46A (3), 533-558.
Garrett, N. (1987). A psycholinguistic perspective on grammar and CALL. In W. F. Smith (Ed.), Modern media in foreign language education: Theory and implementation (pp. 169-196). Lincolnwood, IL: National Textbook Co.
Grace, C. (1998). Personality type, tolerance of ambiguity and vocabulary retention in CALL. CALICO Journal, 15 (1-3), 19-45.
124
Guillory, H. G. (1998). The effects of keyword captions to authentic French video on learner comprehension. CALICO Journal, 15 (1-3), 89-108.
Huckin, T., Haynes, M., & Coady, J. (1993). Second language reading and vocabulary learning. Norwood, NJ: Ablex Publishing Corporation.
Jacobs, A. M., & Grainger, J. (1994). Models of visual word recognition—sampling the state of the art. Journal of Experimental Psychology, 20 (6), 1311-1334.
Kay, A. (1991, September). Computers, networks and education. Scientific American, 138-148.
Levin, J., Levin, M., Glasman, L., & Nordwall, M. (1992). Mnemonic vocabulary instruction: Additional effectiveness evidence. Contemporary Educational Psychology, 17, 156-174.
Livingstone, M. S. (1988, January). Art, illusion and the visual system. Scientific American, 78-85.
Marcel, A. J. (1983). Conscious and unconscious perception: Experiments on visual masking and word recognition. Cognitive Psychology, 15, 197-237.
McCandliss, B. D., Posner, M. I., & Givón, T. (1997). Brain plasticity in learning visual words. Cognitive Psychology, 33, 88-110.
McLuhan, M. (1964). Understanding media. New York: McGraw-Hill.
Myers, D. G. (2001). Psychology (6th ed.). New York: Worth Publishers.
Nagata, N. (1999). The effectiveness of computer-assisted interactive glosses. Foreign Language Annals, 32 (4), 469-478.
Nagy, W. E., & Herman, P. A. (1987). Breadth and depth of vocabulary knowledge: Implications for acquisition and instruction. In M. G. McKeown & M. E. Curtis (Eds.), The nature of vocabulary acquisition (pp. 19-35). Hillsdale, NJ: Erlbaum.
Norman, D. (1994). Things that make us smart. Reading, MA.: Addison-Wesley.
Ong, W. J. (1982). Orality and literacy: The technologizing of the word. New York: Methuen.
Papagno, C., Valentine, T., & Baddeley, A. (1991). Phonological short-term memory and foreign-language vocabulary learning. Journal of Memory and Language, 30, 331-347.
Papagno, C., & Vallar, G. (1995). Verbal short-term memory and vocabulary learning in polyglots. Quarterly Journal of Experimental Psychology, 38A, 98-107.
Paribakht, T. S., & Wesche, M. (1996). Enhancing vocabulary acquisition through reading: A hierarchy of text-related exercise types. Canadian Modern Language Review, 52 (2), 155-177.
Postman, N. (1985). Amusing ourselves to death: Public discourse in the age of show business. London: Methuen.
Raichle, M. E. (1994). Practice-related changes in human brain functional anatomy during nonmotor learning. Cerebral Cortex, 4, 8-26.
Reicher, G. M. (1969). Perceptual recognition as a function of meaningfulness of stimulus material. Journal of Experimental Psychology, 81 (2), 275-280.
125
Service, E. (1992). Phonology, working memory, and foreign-language learning. Quarterly Journal of Experimental Psychology, 45A (1), 21-50.
Service, E., & Kohonoen, V. (1995). Is the relationship between phonological memory and foreign language learning accounted for by vocabulary acquisition? Applied Psycholinguistics, 16, 155-172.
Stoller, F. L., & Grabe, W. (1993). Implications for L2 vocabulary acquisition and instruction from L1 vocabulary research. In T. Huckin, M. Haynes, & J. Coady (Eds.), Second language reading and vocabulary learning (pp. 24-45). Nor-wood, NJ: Ablex Publishing Corporation.
Tufte, E. R. (1997). Visual explanations. Cheshire, CN.: Graphics Press.
Wesche, M., & Paribakht, T. S. (1996). Assessing second language vocabulary knowledge: Breadth versus depth. Canadian Modern Language Review, 53 (1), 13-40.
Zimmerman, C. B. (1997). Do reading and interactive vocabulary instruction make a difference? An empirical study. TESOL Quarterly, 31 (1), 121-140.
AUTHOR'S BIODATA
Lee Forester is Associate Professor of German at Hope College in Holland, MI. He received his Ph.D. from the University of California, Berkeley in Germanic Linguistics. His first book, Umlaut Phenomena in Early New High German: A Pragmatic Approach (Peter Lang, 1999), while not connected directly to CALL, does express the author's keen interest in the history and structure of the German language. He is a member of the Technology Committee of the American Association of Teachers of German and is director of the Auf Geht's! project, a first-year multimedia course in German (Live Oak Multimedia, in development), supported by generous grants from Ameritech and FIPSE.
AUTHOR'S ADDRESS
Lee Forester
Department of Modern and Classical Languages
Hope College
Holland MI 49423
Phone: 616/395-7567
Fax: 616/393-7559
Email: forester@hope.edu
Retrieved on December 20, 2010 from https://www.calico.org/memberBrowse.php?action=article&id=403
Tidak ada komentar:
Posting Komentar