Communication Strategies for Visual Learners

Communication Strategies for Visual Learners

Andy Bondy and Lori Frost

One of the most serious problems for individuals with autism is their great difficulty acquiring and using speech. Almost 80% of preschool students entering one public school program had no functional speech (Bondy & Frost, 1994a). Parents and professionals alike want their students to acquire speech as quickly as possible. The earlier chapters in this book describe an approach that is highly successful in helping these students develop speech. However, for reasons that are not currently understood, some students do not rapidly acquire speech or vocal imitation. When students display significant difficulty acquiring speech, it may be appropriate to consider some form of augmentative or alternative communication (AAC) strategy.

In this chapter, we describe an alternative approach to communication training for students who demonstrate difficulty acquiring a strong vocal imitative repertoire that is, the students who are described in this manual as visual learners. We describe the Picture Exchange Communication System (PECS) and its introduction to individuals with autism and related developmental delays.¹ this system has helped many students learn the basic elements of communication. We also describe the use of pictures and other symbols to help students understand what is expected of them in terms of individual activities and the sequence of activities across a typical day.

What Is Communication?

Before describing several different AAC strategies, it may be helpful to clarify what is meant by communication. First, communication must take place between at least two people, one identified as the speaker, and the other identified as the listener. For example, if a teacher asks a student to bring him or her a hammer, then the teacher is the speaker and the student is the listener. On the other hand, if the student says, 'I want a cookie,' and the teacher brings him one, then the student is the speaker and the teacher is the listener. Communication requires the speaker to behave, or act, directly toward the listener, who in turn rewards the speaker in some manner. Although this description may sound very simple, it is necessary to distinguish communication from other behaviours. For example, a student might go to the refrigerator, open the door, take out a container of milk, and drink some milk. Although this action can be interpreted as indicating that the student wants milk, it will only be confusing to say that the student has communicated his needs. He directs his actions toward the refrigerator and the milk, and not toward any other person. On the other hand, if the student went to someone and said, 'I want milk,' and that person gave the student some milk, then the student's actions may be described appropriately as communicative.

It also is helpful to point out that communication does not require speech; people can learn to communicate without using speech. Individuals who are deaf can learn to use sign language to communicate effectively. Persons who have physical damage to their vocal system (or related physical disabilities) may learn to use various communication devices (some electronic) to type out or select symbols to effectively communicate with others, or they may learn to use a pen to write. In other words, speech is only one modalityalthough clearly the most common and efficientthrough which people can communicate.

One critical element of communication is the reward provided by the listener that supports the communication of the speaker. In the case of teaching students to communicate, it is the teacher who provides the reinforcer for the student's communicative actions. To simplify this analysis, two broad types of reinforcers can be identified as most often involved with communication. One type of reward involves a host of objects, activities, or events that the listener provides to the speaker. For example, when a boy asks for milk, receiving the milk will reinforce asking. In similar fashion, a student may ask for other objects (e.g., toys, magazines, food) or activities (e.g., being tickled, read to, or chased), and receiving these items or activities will promote more communication.

On the other hand, people often communicate for rewards that appear to be purely social in nature. For example, a very young boy may see a dog run by a window and very excitedly say, 'Doggie! Doggie!' The boy is likely to persist until his mother or father says, 'Yes, it is a dog,' or 'Yes, I see the dog, too!' In this case, the boy is not asking for the dog itself and therefore does not need to receive the dog to be rewarded for communicating with his parent. Instead, the reinforcer involves the social interaction (attention) with his parent. For typically developing individuals, this type of motivation appears to be as important (and persistent) in forming early language as the direct receipt of objects or activities.

Even speech imitation can occur for these two different reasons. Most often, children learn to imitate what they hear because of the praise they receive when they can clearly say a word, repeat a phrase, or sing part of a song. On the other hand, sometimes parents (or teachers) may say, 'If you can say 'cookie,' I'll give you the cookie.' In this case, a child may be more likely to imitate because doing so helps to achieve a desired outcomenamely, getting cookies.

When developing AAC strategies for persons with autism, it is helpful to remember that, for many of these individuals, social interactions may not serve as powerful rewards or motivators. In other words, especially early on in training, it is not likely that a student with autism will learn a new skill simply to see his parent or teacher smile. However, during the same period of early training, the student is likely to want a variety of items, activities, or events. If this is the case, then identifying these potential rewards will be important in developing effective communication strategies (see Chapter 7).

What Is Spontaneous Communication?

In addition to understanding the types of rewards that are associated with communication, it is helpful to understand the circumstances (i.e., the stimulus conditions) under which communication occurs. For example, a student may say 'cookie' only when someone models the word or someone specifically asks the student, 'What do you want?' On the other hand, a student might see a cookie, and without anyone saying something, say, 'I want a cookie.' In general, we say that speaking in response to someone else's model or question (i.e., a response to a specific SD provided by another person) is not spontaneous, whereas speaking in response to some aspect of the environment (including things happening within the student, such as feeling hunger or pain) is spontaneous.

A similar line of reasoning can be extended beyond requests. If a student names objects only when asked to do so (i.e., 'What's this?' 'What do you see?') rather than when objects are seen (or heard or felt or tasted), the student's naming of objects would not be described as spontaneous. In general, if a parent's or teacher's model or question is part of what controls the likelihood that the student will respond appropriately, then the student's response is not viewed as being spontaneous. Some students respond only when spoken to; we sometimes refer to these students as being prompt dependent or, to be more technically precise, under narrow stimulus control. (Calling a child prompt dependent misattributes the problem to a supposed characteristic of the child rather than more appropriately noting how the child's behaviour was taught to be dependent upon some teacher-arranged prompt.) These students often need to be taught to use their communication skills in a spontaneous manner because their language does not automatically generalize from prompted communication to spontaneous communication.

What Has Been Tried?

Over the past 20 years, various AAC strategies have been tried with young children with autism. For example, researchers have studied the introduction of manual sign language (Carr, 1982; Reichle, York, & Sigafoos, 1991). As with any AAC strategy, there are advantages and disadvantages to this approach. Among the potential advantages is the strong visual support provided within this strategy and the fact that the model (i.e., the stimulus associated with a teacher demonstrating the appropriate sign) of a particular sign can remain part of a child's visual environment far longer than can the sound of a single spoken word. On the other hand, teaching children how to sign has usually involved first teaching them to imitate the appropriate hand motions. For some children, such imitative skills are acquired very slowly. For those children who do acquire these new skills, there is often a lengthy period of time to acquire functional communication skills. Another potential disadvantage is the limited understanding of sign language in the general population. When someone goes into a fast food restaurant, for example, and has only a sign repertoire, an interpreter is needed (which can limit ultimate independence) or the server needs to understand sign language.

Another general approach to AAC involves the use of pictures and related symbols. Until recently, these strategies involved pointing (or similar selection strategies) to pictures or symbols that correspond (or match) various items or events. An advantage of a picture-point system is the fact that there is no need for the student to imitate someone's actions (or words) prior to being taught to use the pictures. Furthermore, some students appear to attend more consistently to visual cues than to auditory cues; such students may then be more attentive to pictures than to spoken words. Pictures also are readily understood by naive communicative partners, such as people in the community.

On the other hand, several potential disadvantages are associated with attempting to teach a picture-point system as a means of teaching functional communication. With these systems, early lessons are led by the teacher²; that is, the teacher begins the interaction by holding up an item; asking, 'What's this?'; and rewarding the student if she points to the corresponding picture. Some students appear to learn to point to pictures only when asked to do so and do not initiate communication interactions. Other students have been observed to be sitting in a corner, tapping on a picture, and waiting for someone to notice what they are doing. In other words, these students may not have learned to approach someone prior to pointing to the picture, and thus a key element in basic communication has not been acquired. Another difficulty occurs when a student points to a picture but simultaneously looks away. In such cases, it is difficult to interpret the pointing: does the student want what she is pointing to? Does the student want what she is looking at? Does the student simply like the sound of tapping on the picture? Such reasonable questions of interpretation may severely interfere with the development of functional communication skills. Throughout these examples, it is assumed that the point of the lesson is to teach the student to be able to engage in appropriate and self-initiated functional communication rather than to test whether the student can match to sample. That is, although matching to sample is an extremely important skill, it does not necessarily automatically generalize to spontaneous requests or comments.

What Is the Picture Exchange Communication System?

The Picture Exchange Communication System (PECS) was developed within the Delaware Autistic Program during the past decade (Bondy & Frost, 1994b).³ The primary purpose of introducing PECS to students is to teach them to initiate a communicative interaction within a social context (i.e., approach an adult and deliver a message). From the start of training, the student is taught (without depending on verbal prompts) to approach someone who has something (e.g., a favourite snack or toy) that the student clearly wants. This communication function is taught first because it results in immediate reinforcement for the student and does not rely upon social rewards that may be ineffective at this point in training. The system begins with using single pictures within a request function, then develops symbol selection (i.e., discrimination), then develops sentence structure, and finally addresses additional communicative functions.

Assessing What a Student Likes

Learning about a student's preferences involves doing a reinforcer assessment. Such assessment can be very formal (i.e., highly systematic with very controlled arrangement and data collection); most often, however, we suggest more informal strategies, especially for parents. Most parents already have a very good understanding of what their child likes (and dislikes). Teachers and other professionals should always start the process by discussing this issue with parents. The process may begin by looking for items for which the student readily reaches without direct prompting. If the student reaches for an item, the teacher may try to see how motivating the item is by making access slightly more difficult. For example, the teacher could grasp an item until the student pries open his fingers, or the teacher could put the item on a low shelf and observe if the student still reaches for the item. This rein-forcer assessment should continue until the teacher identifies several items (e.g., 5 to 10 toys or snacks) the student desires.

Various formal and informal strategies are associated with determining what is reinforcing. Among the issues that must be addressed within any strategy are factors that may lead to satiation (i.e., offering a reinforcer so frequently that its potency is seriously, though temporarily, diminished) and the number of potential reinforcers offered (i.e., presenting only one item and observing how hard the student tries to obtain that item vs. offering a choice of reinforcers).

Phase 1: Beginning PECS Training

After the desires of the student have been determined, then PECS training can begin. The first step of PECS is to teach a student the essence of communication, that is, to approach an adult rather than to act directly on an object. (Issues associated with selecting particular messages will be addressed during subsequent lessons.) Therefore, the first element of PECS training involves teaching a student to pick up a picture, extend the picture to the teacher (who functions as the communicative partner for the student), and release the picture into the open hand of that person. Of course, the teacher immediately gives the student what was requested. Although this sequence sounds very simple, this training step requires a special arrangement; that is, the training requires that two teachers work with the student. Arranging for one-on-one instruction may seem complicated enough, and planning for two teachers may seem daunting. However, there is need for two teachers for only a short period of time. Attempting to start PECS with only one teacher will likely lead to long-term problems.

Training begins with one teacherthe communicative partner or the 'enticer' sitting in front of the student and displaying something the student wants. One role for this teacher is to entice the student to reach for the offered item. For this example, we will assume that the reinforcer assessment indicated that the student likes cookies. A picture of a cookie is placed between the student and the cookie. The enticer neither asks the student what she wants nor tells the student to pick up the picture. In fact, this teacher should not use any direct verbal cues to help the student learn to initiate the interaction.

The enticer may put out one hand, palm up, to accentuate his or her role as a receiver. (This gesture will need to be eliminated over time.) The second adultthe physical prompter or the 'helper'sits (or stands) immediately behind the student. This teacher must wait for the student to begin to reach for the cookie. If the helper tries to move the student's hand toward the picture before the student is actively seeking the desired item, then the student may resist the attempt to move her arm. As soon as the student moves toward the cookie (and thus is orienting to the reinforcer), the helper guides the student's hand to pick up the picture, extend it to the open hand of the enticer, and release the picture. As soon as the picture is received, the enticer says, 'Cookie! You want the cookie!' in an animated fashion and gives the student the cookie. (An alternative statement by the enticer could be 'I want cookie.' This phrase immediately models the correct form for the child to eventually imitate. At this time, there is no clear empirical evidence to prefer one type of comment over the other.)

The reason that two teachers are necessary is that there are two separate roles for these teachers from the student's perspective. One teacherthe enticerhas the item that the student wants. The student's initial tendency, as observed during the reinforcer assessment, is to reach for the item and take it. Because the student will be taught to pick up a picture and give it to the enticer, someone must help the student to pick up the picture, extend it to the partner, and release the picture into the enticer's open hand. The teacher helping the student should not be the enticer. If the enticer helps the student, then the student may learn to depend on that help and not learn to initiate the exchange.

The role of the helper is very different. His or her first goal is to wait for the student to indicate, by an action, that the object is desired. The most common action to observe is the student reaching for the object. As soon as this action is observed, the helper quickly prompts the student by physically guiding the student's hand to the picture, helping the student (with hand-over-hand assistance) to pick up the picture, extend the picture to the open hand of the enticer, and release the picture into the enticer's j open hand. Notice that this second teacher does not say j anything to the student before or after the exchange. In j fact, the student should show very little awareness of this! Teacher and instead focus attention on the desired object, the picture, and the enticer's open hand.

It is important to remember that the physical prompter helps the student only after the student starts to reach for the object (or for the picture in later trials). The prompt for the student to pick up the picture should never begin while the student is looking out the window, at her own fingers, and so on. At such times, guiding the student to the picture will have no meaning for the student. If the student seems distracted, the enticer (not the helper) should attempt to attract the student's attention to the potential reward. Once this attention is gained, the enticer can move the object, nibble on the snack, even comment on how good the snack tastes. Even with these enticements, the enticer must still avoid directly telling the student to pick up the picture or asking the student what she wants. The use of such direct prompts may teach the student to wait for the prompt instead of learning to initiate the communicative interaction with someone who has what is desired.

The primary responsibility of the helper is to fade the level of support as rapidly as possible. The most effective way to fade support is to take the view that the student has three steps to learn: (1) picking up the picture, (2) extending the picture toward the enticer, and (3) releasing the picture. Fading should take place through the use of backward chaining (see Chapter 10). To use this strategy as it relates to PECS, support is provided for the first component in the sequence and reduced for the last step. Thus, the helper assists the student in picking up the picture (Step 1) and extending the picture (Step 2), and then fades assistance in releasing the picture (Step 3). Once the student releases the picture without help, the helper guides the student to pick up the picture and begins to fade help in extending the picture toward the enticer. Finally, assistance is reduced for helping the student pick up the picture.

The main point of this first phase of training in PECS is to teach the student to initiate an action toward the picture and proceed to give the picture to the teacher. Many students learn this step within the first session of training (i.e., within the first hour of training). One potential area of difficulty is associated with very low motivation by the student toward the possible reward. To minimize this potential problem, it is helpful to assure that the student continues to reach for freely offered items. If the student does not display much effort in obtaining the reward, she will not display more effort to pick up a picture to exchange for the reward.

It may also be helpful to remember that, although it is preferable to work with the student while seated in a chair, this lesson may need to be arranged in other ways to assure that the student is adequately motivated by what is offered. For example, a lesson can be arranged while the student is playing with blocks or puzzle pieces on the floor. These items are motivating to the student if she is persistent in picking up additional blocks or puzzle pieces. In this situation, the enticer should hold the next puzzle piece and place a picture of a puzzle piece in front of the student. Once the student reaches for the puzzle piece, the helper then assists the student in initiating the exchange as described earlier. Essentially, this lesson should be arranged where the most potent rewards can be controlled by the enticer.

Although the first phase of training may be accomplished with many types of rewards, we have found certain types of rewards easier to use than others. First, we use very small portions of any reward to promote as many rapid trials as possible (and thus reduce satiation). We also prefer to start with rewards that tend to be consumed very quickly. For example, snacks and drinks are swallowed rapidly, bubbles tend to pop quickly, spinning tops tend to stop moving relatively quickly, and radios or tape recorders can be readily turned off. Other types of rewards are often controlled by the student and are difficult to end. For example, looking at a book or playing with a toy may be difficult to stop (i.e., to create another opportunity to request) because the student may get upset when the object is taken by the adult. Although we try to avoid such rewards in the beginning of training, if we need to use these rewards, we simply announce 'My turn' and calmly take the object away, rather than trying to begin a 'Give me ' lesson. If the student is mildly upset by the removal of a favourite object, that is, in our opinion, a good indicator to continue the lesson. We ignore the mild negative reactions and continue to try to guide the student to perform correctly. Of course, if this strategy results in dangerous behaviours (either toward the teacher or self-inflicted), then we suggest using another reinforcer. The goal is to gradually teach the student to request many different types of reinforcers. In our experience, the order in which they are taught to be requested is not critical.

To avoid satiation, it also may be helpful to rotate among different types of rewards, each individually presented with its corresponding picture (see the next section for details). Furthermore, to help maintain motivation over time, keep each of these early lessons relatively brief, 5 to 15 minutes at most. At the same time, disperse this lesson across at least half a dozen opportunities per day. The more the student appears to want specific reinforcers, the more opportunities there are to work on requesting.

During this initial phase of training (and during the next phase, as well), the student is not given a choice of pictures. It is the teacher's responsibility to focus on the importance of the exchange. In our experience, some students appear to look at and manipulate the pictures in a manner that suggests that they know or recognize what the picture represents. On the other hand, some students do not appear to look closely at the contents of the picture. This issueselection and discrimination of picturesis addressed in the third phase of training.

Phase 2: Increasing Spontaneity

Once the student successfully completes the first phase of training, he can independently pick up a single picture and give it to the teacher immediately in front of him in exchange for a desired item held (or controlled) by the teacher. (We consider the child to have successfully completed the first phase when he initiates and completes a request without a prompt from the helper on at least 80% of opportunities presented in several situations across the day.) Now the student is ready to learn the steps in the next phase of training. Phase 2 will expand the types of rewards the student can request, increase the distance the student walks to the teacher, increase the distance between the student and the picture, and introduce several different teachers with whom the student can communicate.

To accomplish the goals in Phase 2, each aspect is individually addressed. During this phase, it also is important not to place two (or more) pictures at the same time before the student. Discrimination between pictures will be addressed in the next phase of training. First, how to introduce additional types of rewards within the PECS framework is described.

In general, Phase 1 of training is often accomplished using only one type of reward. It is helpful to avoid associating communication with only one type of reward, such as snacks, so that the student does not come to expect that all pictures bring food. Rather, the goal is to help the student learn that giving pictures is a way to communicate with other people about many different things that the student wants. The second type of reward should be something distinctly different from the first reward. For example, if the first training item involved pretzels, then the next object may involve puzzle pieces or bubbles (depending on the student's preferences) rather than another type of snack food.

It is also important to assure that the student will use the system with many different adults (and ultimately peers). Therefore, very quickly, even during the first training session, the enticer and the helper should switch roles. Then other people, such as other staff and family members, should be included. The key is to be certain that each person with whom the student communicates knows the types of cues to use (e.g., enhancing the potential reward) and the type of prompts to avoid (e.g., direct questions or commands, continued use of an open hand). The continuity in strategy across different teachers will help the student generalize this new skill. We have observed little difficulty in teaching students to use the system with other people because they tend to be focused on obtaining the desired objects and the necessary pictures.

Another aspect of increasing generalization involves having the student learn to move around the room with his picture by gradually increasing the effort needed to complete the exchange. First, the enticer begins to sit straight or even leans back in his or her chair, as opposed to leaning forward, as is usually the case in Phase 1. Then, the enticer gradually moves away from the student, literally inch by inch. Some students show a tendency to remain in their chair; in such cases, the helper, rather than the enticer, should prompt the student with the minimum guidance necessary to start moving toward the partner. Usually, the combined outcome of getting the reward and hearing an enthusiastic reply by the enticer assures subsequent approaches.

Another strategy to further enhance generalization (or independence) involves teaching the student to get his own pictures. This lesson is made easier to teach by placing each picture on a communication booklet with Velcro. Every student using PECS should have his own communication booklet. Three-ring binders have been used with good success. At the earlier phases of PECS, the picture or pictures that the student are currently using are placed on the cover of the binder and other pictures (ones that could be used in other situations) are kept on pages inside the binder. The student learns to pull the picture from its Velcro attachment. This system also helps to minimize losing the pictures. At the start of this phase, the binder is placed immediately in front of the student with only one picture showing. Gradually, the binder is moved away from the student so that the student must move to the binder to retrieve each picture to exchange with the teacher. One helpful option is to maintain the binder in one easy-to-reach location within the home (or group situations), such as attached (via a loop or Velcro) to one particular area on a wall or on a specific chair.

One successful strategy to teach a student to independently retrieve his own communication book involves periodically entering a room eating something that the student likes, such as a bowl filled with popcorn. Start with items that the student has already learned to request in controlled situations. Furthermore, make sure that the corresponding picture is on the front of the student's communication book. If the student simply walks over and reaches for the snack, ignore him and continue to slowly walk around the room while eating the food. If the student does not quickly go over to his communication book, casually wander over to the area where the book is kept. Once the student retrieves his book and offers the appropriate picture, share some popcorn with him. The picture should then casually be returned to the book, and a new trial should begin by walking a little farther from where the book is kept. Over time, students learn to look for their communication book rather than first grab for something that they want. In our opinion, it is helpful to arrange these lessons so that they do not appear to be lessons.

When the student successfully completes this phase of training (by independently initiating in at least 80% of opportunities presented in multiple settings across the day), he can initiate requesting a variety of rewards (individually presented) by retrieving a single picture from his communication binder from anywhere in the teaching area and with a variety of teachers. In general, students typically use from 6 to 12 pictures in three or four different situations with two to four people by the end of this phase.

Phase 3: Discrimination Training

During the initial phases of training, it is possible for a student to select a single picture from the communication board without actually understanding the relationship between the picture and the corresponding item. That is, the student may actually not have learned to 'look' at the contents of the picture. This next phase of PECS training focuses on this important issue. As indicated in other chapters of this book, there are several strategies to approaching discrimination training. Rather than reviewing all of these strategies, several will be described that have been helpful with many students. It is also important to note that research on discrimination strategies, including when to begin discrimination training, is ongoing, and therefore much of what follows is best understood as being in accordance with clinical practice, as opposed to derived from specific studies on each topic. (In our opinion, it is possible to begin discrimination training within the PECS format prior to beginning matching-to-sample training. Therefore, we do not recommend waiting to introduce PECS (or its discrimination phase) until a child has successfully completed matching to sample on two-dimensional representations. More research is clearly needed in this important area of training.)

Discrimination training can begin with one teacher and is started within a situation familiar to the student requesting desired items. The first approach involves presenting the student with two items and two corresponding pictures. One of the items should be something that the student finds very rewarding and the other item should be something that the student simply does not seek to acquire. This pairing creates a situation where there is a big contrast between the items that the student can choose. For example, a favourite food item may be presented along with a sock (assuming the student is familiar with socks but does not seem to want one). Both items are displayed or offered to the student. To be certain that the student is looking at the two choices, training begins after the student clearly has demonstrated a preference for one of the actual objects. The teacher does not need to use any type of verbal prompt at this point (i.e., by asking, 'Which one do you want?' or 'What do you want?'). Next, the pictures are immediately presented to the student. If the student selects the snack picture, as soon as the student touches the picture of the snack, the teacher says 'Oh' in an upbeat fashion. When the student then gives the snack picture to the teacher, the student immediately is given the snack.

If, on the other hand, the student reaches for the picture of the sock, as soon as she touches that picture, the teacher would say something like 'Uh-oh' and, upon receiving the picture, give the student the sock. (We suggest that the teacher use a neutral tone and phrase to minimize the possible reinforcing effect that teacher-provided feedback may provide. We have observed situations where the tone of voice used following selection of the 'wrong' picture was so reinforcing in itself that subsequent discrimination was very difficult until the tone was modified to reduce its reinforcing association.) At this point, most students look at the sock and then look at the teacher as if the teacher must be crazy. Some students simply get upset, and toss the unwanted item or begin to fuss. If this happens, that is a good sign because the outcome of the interaction with the teachergetting a sockdoes not fit the student's expectation. The situation is quickly repeated, even for a student who is clearly upset about not getting what she expected, by re-presenting both pictures, and the correct picture selection is immediately reinforced. Over time, the location of the pictures relative to each other should be varied to prevent position cues that interfere with attention to the visual contents of the pictures.

It is important to provide some appropriate verbal feedback the moment the student touches a picture (the earliest indication of a selection) rather than wait for the student to give you the picture. (Even if such comments are not initially effective reinforcers, such comments should come to function as conditioned reinforcers over time given their association with the receipt of established reinforcers.) If you wait for the exchange, the feedback comes too late (i.e., it is provided after the picture is given, not after the picture is selected) relative to when the student actually made the choice between pictures. How-ever, if the student is given the sock and simply begins to play with the sock without seeming to be upset that she did not get the snack item, then this strategy may not be very successful. Simply stated, if the student does not seem to care about the outcome of a request, then the student is not very likely to be motivated enough to study the pictures prior to making a choice. Discrimination strategies that yield many errors tend to be less successful than those that can minimize the frequency of errors during acquisition.

If this strategy does not work, you may want to try the following strategies:

Spatially align the pictures with their corresponding items. Gradually increase the distance between the pictures and their corresponding items.

Put the pictures on clear containers holding desired items; over time, make the containers more and more opaque so that the student must 'see' the picture to understand what is in the container.

Enhance the visual differences between the pictures by, for example, (a) using a picture of a desired item and a 'blank' distracter card (i.e., a card the same size as the picture of the desired item but without a picture on it); (b) initially making the picture of the desired item larger than the picture of the nondesired item, and then gradually making the pictures the same size; (c) initially enhancing the colour of the picture of the desired item by colouring the picture or placing a colour frame around it, and then gradually reducing the amount of the colour enhancement; (d) using product logos, photographs, or other enhanced visual symbols (including the small pictures found on most coupons of favourite foods and drinks) rather than black-and-white line drawings.

4- Use three-dimensional objects instead of pictures. Plastic replicas can be used for many items (often found as refrigerator magnets), or real items can be covered with an acrylic epoxy (found in most art supply stores). Over time, the encased objects can be gradually covered by pictures (Frost & Scholefield, 1996). (See Frost & Bondy, 1994, for more details on these and other discrimination strategies.)

Another important strategy to introduce during this aspect of training is correspondence training. In general, when a student selects a picture of a pretzel from a pair also including a picture of a sock, the teacher would immediately say, 'You want the pretzel! Here it is!' and give the student the pretzel. However, if the choice was between a pretzel and a potato chip, the student may not strongly prefer one over the other, as long as she gets something. To determine whether the student is actually making a true selection, after the student gives the teacher a picture (from the pair presented), the teacher should say, 'Okay. Go ahead. Take one.' Notice that in this case, the teacher would neither say the name of the requested item nor give the item to the student. Instead, the student is simply offered the choice between the items available. If your student reaches for and takes the item requested, say, 'Good. Nice paying attention,' or something to that effect. However, if the student reaches for the other item (i.e., she gives you the picture of a pretzel but selects a potato chip), then do not permit the student to take the item. Remind the student what she asked for and demonstrate the association between the picture and the item. Set the situation up again (with both pictures), and reward the student for appropriately following up on her picture choice by selecting the corresponding item. (A more comprehensive error-correction sequence that is appropriate in this situation, and for related errors, can be found in Bondy & Frost, 1995.)

After the student masters two-way discrimination between pictures, the number of pictures from which the student can choose is gradually increased. At first, pictures of items the student does not really want are introduced (e.g., one picture of a desired item and two pictures of nondesired items). As the student's performance improves, pictures of things the student may want but that do not fit the current situation are gradually added. For example, while the student is watching TV, the teacher might turn the TV off and present the communication book with a picture of the TV and a picture of a spoon. In this situation, the student is likely to want the TV. At another time, for example, while the student is eating her favourite dessert (e.g., a bowl of ice cream), the same picture choices may be offered. This time, the student is more likely to request the spoon.

The criteria for moving on to the next phase of training involve both the accuracy of the student's performance and the number of pictures in the student's repertoire. Successful discrimination (80%) within an array of at j least 5 pictures (arranged as an X) should be demonstrated over a period of at least 3 days. The total repertoire should be at least 10 pictures. At the same time, it is j important to foster moving on to the next phase rather j than simply expanding the pictures used within this stage   (i.e., such as teaching the discrimination of 50 single pictures without developing sentence structure).

Introducing Sentence Structure with PECS

Up to this point in the student's use of PECS, individual pictures have been manipulated. This level of communication development can be compared to that of typically developing individuals who can use only single spoken words. However, typically developing individuals can convey both requesting and labelling functions while using single words by modulating the tone of their voice (or other aspects of their word production). For example, an 18-month-old can voice 'doggie' in a manner that conveys that he wants his favourite dog toy or in a manner that implies that he sees a dog running in the yard. Students using PECS will need a comparable mechanism to express whether the picture they are exchanging is used as a request or a comment (label). To accomplish this communicative function, we teach students to develop simple sentences using more than one picture.

To introduce this phase of training, create a removable card (a sentence strip, which is typically about 5x2 inches) and an additional card that has 'I want' written on it as well as an accompanying symbol (e.g., a pair of open hands). The phrase 'I want' is presented as unitary because the student has not yet learned to distinguish 'I' from 'you' or the names of other people; that step will come later in training. Training begins with the 'I want' card already placed on the sentence strip so the lesson that the student first learns is to place the picture of what he wants on the sentence strip and then to give the sentence strip (instead of only the single picture) to the teacher. As in the previous phase, there is no need for verbal prompting by the teacher. Physical prompts are used after the student begins to hand over the single picture to guide his hand toward placing the picture on the sentence strip. When the teacher receives the sentence strip, in addition to giving the student what was asked for, the teacher reads back the sentence strip while pointing to each picture (i.e., 'Oh, 'I want' [pausing 2 or 3 seconds] cookie!'). The pause before naming the item requested tends to encourage students to say the name of the item. (The critical word here is encourage. The child is not forced to repeat any of the words in this lesson. The child performs correctly by appropriate manipulation of the pictures and thus deserves a reward for learning to use the sentence strip.) Once the student learns to place the single picture on the sentence strip and give the entire sentence to the teacher, the student is taught to take the 'I want' card and place it on the sentence strip (using backward chaining).

Answering Simple Questions

So far, the student has learned to use the PECS Program to spontaneously request a variety of items. In addition to using spontaneous communication, it is important for students to respond to questions. The next step in the PECS sequence introduces the question 'What do you want?' Although teachers have been encouraged to avoid that question up to this point in training, the student likely has heard the question. In either case, this step in training usually proceeds relatively quickly. This step also introduces a prompting strategy that is used in subsequent phases.

To begin this phase of training, an item that is reinforcing to the student is placed before the student, along with her communication book. While asking, 'What do you want?' the teacher simultaneously points to the 'I want' picture on the student's communication book. Most students will pick up the picture to which the teacher points and proceed to make a request with the pictures. Over time, a pause is introduced between the question and pointing to the 'I want' picture (this strategy is described as a progressive time-delay prompt procedure). As the delay is gradually increased, the student usually begins to respond to the question before the teacher points to the 'I want' picture. Thereafter, the teacher can eliminate the gestural prompt and promote independent answering of the question. Of course, it is still important to arrange for the student to have opportunities to respond spontaneously as well as be able to answer this question.

Teaching Labelling in Response to Simple Questions

At this point in training, the student can request items spontaneously and respond to simple questions. Criteria for moving on to the next step, labelling objects and items in response to simple questions, would be observing the student successfully respond to the question, 'What do you want?' in at least 80% of no less than 20 opportunities presented over 3 days. The new labelling skill is taught by relying upon previously learned skills. A new picture is added to the communication board'I see.' This lesson is begun by placing before the student an item that, although somewhat interesting, is not something that the student strongly desires. The teacher asks the student, 'Look! What do you see?' while simultaneously pointing to the 'I see' picture. This gestural prompt was introduced in the previous phase; therefore, the student is likely to pick up the 'I see' picture, place it on the sentence strip, and select the picture corresponding to the item before the student. Upon receiving the sentence strip, the teacher should say, 'Oh, I see the ball.'

At this point it is very important not to give the ball to the student. The appropriate reaction to a student's naming or commenting about something is to agree with the student and provide social reinforcement. Naming an item is not the same as requesting an item. Therefore, to avoid confusing the student, the teacher must react differently to requesting than to naming. The reason that an item that is not highly reinforcing for the student is selected is to lessen the reaction the student may have when he does not receive the item, something the student most likely expects given his prior history with PECS. Over trials, the teacher begins to introduce a pause between asking the question and pointing to the '1 see' picture (as done in the previous phase). As the delay is increased, the student gradually comes to respond to the question without any gestural prompt.

During this phase of training, the teacher introduces other simple questions, such as 'What do you hear?' 'What do you feel?' and 'What do you have?' For each question, a corresponding picture is added to the communication book. Of course, it is very important to continue to mix in questions such as 'What do you want?' along with providing the opportunity to maintain a reasonable rate of spontaneous requests.

Developing Spontaneous Comments

In addition to naming items when asked to do so, it is important for the student to spontaneously comment about aspects of the environment. For the student to take this step, questions by the teacher must be eliminated as prompts for the student's comments. One strategy for accomplishing this involves fading the length of the teacher's questions by, for example, saying, 'Look! What do you ?', then 'Look! What ?', then 'Look! and then using just an expectant look. Comparable phrases can be used with listening (i.e., 'I hear'), feeling or touching (i.e., 'I feel'), or simply holding other tems (i.e., 'I have '). In addition, it is helpful to promote spontaneous comments by using surprise or novelty to gain the student's attention. Typically developing individuals tend to comment about aspects of the environment that suddenly change (as when someone walks into a room, something falls or breaks, something does not 'fit' into a situation, etc.) rather than static aspects of the environment (e.g., the wall or the floor). Therefore, one helpful strategy is introducing change into the environment by, for example, taking a variety of items out of a bag or playing different sounds on a tape recorder.

Introducing Attributes and Other Complex Aspects of Communication

Another aspect of complex communication involves the description of various attributes associated with objects or events. These attributes can include colour, size, shape, type, brand, and so on. These attributes are sometimes first introduced when teaching a student to appropriately respond to teacher requests, such as when asking a student to give the red ball, point to the square, or find the chocolate cookie. On the other hand, once a student learns to use PECS to request items with a sentence strip, it is possible to teach the student to use certain attributes within the request prior to using that attribute within a receptive task.

Assume, for example, that a student not only likes M&Ms but prefers red ones (this lesson could be arranged for a student who prefers Oreos over other cookies, Power Rangers over other figurines, etc.). When the student requests an M&.M by giving the teacher the sentence strip, the teacher holds up several pieces of candy and asks (or simply gestures), 'Which one?' The teacher now guides the student to select a picture signifying 'red' and guides the student to place that symbol on the sentence strip (between 'I want' and 'M&M'). The teacher then clears the sentence strip and guides the student through the new sequence. When the sentence strip including 'red' is received, the teacher gives the student the red M&M. No additional reinforcer is needed because the student specified it (although it is always a good idea to add praise or other types of social reward). Another strategy used to assess whether the student truly wants what is indicated on the sentence strip involves having the teacher simply say, 'Okay. Take it!' when handed the sentence strip. Rather than immediately giving the student the specified item, the teacher should observe whether the student chooses the item that corresponds to what was requested.

This lesson proceeds to other attributes as long as the teacher can determine how to make the attribute an aspect associated with what the student wants. At times, this approach may be simple, as when teaching a student to ask for the long versus short pretzel or the big versus small cookie, because most students prefer the longer pretzel or the bigger cookie without any specific training. (Although the child selects the long over the short pretzel without training, she needs to learn how to communicate about such choices.) Selecting body parts may be made rewarding by first teaching the student to play a game such as Mr. Potato Head and then teaching the student to request the body part that she wants to place next on the body. In general, an attribute must first be potentially reinforcing to the student before the teacher tries to teach the student to ask for the item using that attribute. For example, if a student never displays any preference along the dimension of colour (i.e., the student never chooses favourites by colour), then it will be initially difficult to have the student respond to that attribute within a request. In our experience, lessons involving requesting using attributes to which the student already attends are usually more motivating to the student than lessons based on the teacher's selection of attributes and reinforcers.

The Codevelopment of Speech During PECS Use

PECS has been taught to a large number of children and adults with autism and related delays (Bondy & Frost, 1994a). One change that has been reliably observed when using the system with preschool-aged students has been the codevelopment of speech. For some students, speech has been heard within a few months of training; for other students, this development has not been observed for 1 or 2 years; and for another group of students, speech is never reliably displayed. (There are no data yet, however, concerning the codevelopment of speech following the introduction of PECS for children who have previously had extensive, though unsuccessful, training in vocal imitation. The data presented involve children for whom PECS was introduced almost immediately upon their entrance into a school program.) When preschool-aged students with autism develop speech after the introduction of PECS, these students usually go through a period of training when they communicate solely via PECS, followed by a period of time during which they begin to speak while using PECS, and then followed by their complete reliance upon speech as their mode of communication. Although this development has been observed in any (but not all) students, it should be made clear that PECS is taught in order for the student to have a successful means of communicating and not as a method of teaching speech. Several factors seem to be associated with the development of speech.

First, the introduction of PECS to students does not diminish the importance of working on the development of imitation, including vocal imitation. In addition, within the PECS routines, when the student gives the teacher a picture or a sentence strip, the teacher says the appropriate statement back to the student (thus providing a vocal model), including pausing before saying the last word (or phrase). In our experience, many of the students who begin to speak begin to complete the teacher's statement or begin to say the words with the teacher while pointing to each picture. For the students who develop speech, there is a period of time during which they are likely to speak only while using their picture system. These students often use more complex spoken language (i.e., use more words or involve more communicative functions) when they have the opportunity to use PECS (and simultaneously speak) than when they have no pictures to manipulate (Frost, Daily, & Bondy, 1997). During this initial period of vocal acquisition, most students continue to acquire new vocabularies via PECS before they stop using their pictures and rely solely upon spoken language. Although it is tempting to immediately remove all pictures upon hearing the child's first spoken words, the behaviour patterns of many students suggests that too early a removal of the pictures can arrest the communication development of some students. Of course, all spoken language will be strongly encouraged and reinforced as it develops.

Not all students who use PECS will develop speech. Research is currently being conducted to help predict which students are likely to learn to speak, but we do not have an accurate early predictor at this time. However, we stress that PECS is taught as a way to provide the student with functional communication and not as the primary way to get a student to speak.

Use of Pictures and Symbols Within Receptive Language Lessons

As noted at the start of this chapter and throughout this book, some students display considerable difficulty acquiring speech. In similar fashion, some students have difficulty understanding the spoken language of others (see Peterson, Bondy, Vincent, & Finnegan, 1995, for an example of this difficulty). For these students, their performance in response to teacher directions (e.g., 'Get the spoon') significantly improves when visual cues are added to the spoken instruction.

In some situations, the teacher may use a picture of an item, such as a drawing or photograph of a spoon, while saying, 'Get the spoon.' In other situations, it may be more effective to present an object associated with a particular activity, such as by giving a student a ball while saying, 'Go to the gym.' There is no simple test to determine beforehand whether pictures or more concrete objects will be effective with a given student. Some students may need to learn to discriminate between actual objects or 3-D representations before learning how to discriminate between black-and-white line drawings (Frost & Scholefield, 1996). Some students improve not only in their ability to understand what is said to them, but also show a reduction in disruptive reactions during instructional routines, such as fewer tantrums, self-injurious behaviours, or aggressive acts (Peterson et al., 1995).

When such visual cues are added to a teacher's instructions, the teacher should observe whether the student can eventually successfully respond to spoken instructions. Once the student masters successfully responding to the presentation of the visual cues, the teacher begins to reduce the presence of the visual cues. One such strategy can be introduced by gradually reducing the length of time the visual cue is displayed. For example, early in training, the picture is shown to the student during the entire time between the instruction and the student's retrieving the requested item. Over time, the teacher shows the picture for shorter and shorter periods5 seconds, then 3, then 1, and then just a flashbefore removing it entirely. Another strategy is to gradually introduce a pause between saying the instruction 'Get the spoon' and showing the picture of the spoon. As this time delay is increased, some students begin to respond to the spoken instruction before the picture cue is introduced. Research associated with this time delay prompting strategy is ongoing (Sulzer-Azaroff & Mayer, 1991).

These strategies can be used when using more concrete objects, especially during transitions from one activity to another or one location to another. Using objects associated with the next activity may help students focus on what the next reinforcer will be (e.g., 'Here's your crayon; let's go inside and colour') rather than on the current reinforcer they are losing (e.g., 'Stop playing and go inside'). Such items can be termed transitional objects (Kanter, 1992) as they may assist the student in understanding what activities are about to begin.

Following a Schedule

Many adults use notes, personal calendars, and other visual aids to remember daily or weekly responsibilities. Many students will be able to move more smoothly through the activities of their day if they are taught to use similar visual cues. When a student is taught to follow certain instructions provided as pictures or other symbols, he is taught to respond to the pictures by retrieving some material and then beginning a particular activity. Schedule following involves teaching students to follow a sequence of pictures. These pictures can be presented in a vertical array (or a left-right sequence as the student gets older) or within a booklet, whereby each page has one picture associated with one activity.

The key to teaching successful schedule following is to minimize the teacher's verbal prompts so that the student comes to rely upon the schedule symbols to get the material, start an activity, complete the activity, put the material away, and return to the schedule array to continue with the sequence (Macduff, Krantz, & McClanna-han, 1993). In general, if the student must be told, 'Check your schedule,' then independence will be harder to achieve than if the teacher uses more physical guidance (which is faded over time).

An advantage of using visual schedule systems with students is that the teacher can review the activities and expectations of the day with the student, potentially reducing surprises the student may find disconcerting. The visual schedule also permits pointing out when the most reinforcing activities will occur. Posted schedules also permit some degree of negotiation with a student. For example, a student may request going to a fast-food restaurant at 9:00 in the morning. Rather than saying, 'No,' a parent can choose to put the picture of the restaurant on the schedule at a more appropriate time (i.e., lunchtime or dinnertime). In this manner, the student begins to understand that, rather than being totally rejected, he will have to wait to get what is wanted.

Another advantage of using schedule systems is that the teacher can include additional information in the picture. For example, a card on the schedule can indicate that the student will be starting an art project. On the back of the card (or along the bottom edge of the front of the card), the teacher can place pictures associated with each of the materials needed for the project (e.g., paper, crayons, scissors, paste). This system can become very useful for older students who are learning to complete vocational activities that include many items or for students who are learning to follow cooking menus that include many ingredients and utensils.

Pictures and Reinforcement Systems

As has been stressed throughout this book, providing appropriate reinforcement is the backbone of behavioural approaches to teaching. The first part of PECS training relies upon the student learning to request desired reinforcers. In the beginning of such training, it is important to reinforce the request immediately and on a consistent basis. However, it should be understood that a student cannot be given everything that is requested at all times. For a student who has never calmly requested anything, during the first few days of PECS training, when that student repeatedly asks for a piece of pretzel, it is certainly recommended that each request be rewarded. However, as time goes by, no one will want to (or at times be able to) give the student a pretzel every time one is requested. The student needs to continue to be rewarded for making requests; however, teachers and parents want to make it clear that they gradually expect more and more from the student.

One potential solution involves teaching the student to use a visual reinforcement system. The sequence is similar to how all individuals are taught to value money; that is, people learn to appreciate and desire money because they learn that they can exchange money for things that they want. For a student who uses PECS, the process begins when she makes a specific request. For example, when the student requests a pretzel, the teacher places the picture (with Velcro or some similar material) on a special card (about the size of an index card) while saying, 'Oh, you want the pretzel.' The teacher places on this card a small circle within which is placed (also with Velcro) a type of symbol or token (e.g., a penny, or more age-appropriate trinket or sticker). The teacher immediately asks the student to do a simple task, such as retrieving a pencil on the table or putting a toy in a box. Upon task completion, the teacher gives the student the token, guides her to place it on the card, and immediately takes the token from the card while giving the student what she requested. (The student is not required to give the teacher the picture because she already requested the pretzel and does not need to ask again.)

The student learns the value of the token by trading it in for the desired item. Additional circles are gradually placed on the card until the student needs to earn five tokens before trading them in for the item requested. In this manner, the student has a continuous visual reminder of what she is working for, while at the same time she is gradually taught to accomplish more (i.e., extend the length of time or number of items needed to complete certain lessons or other activities) before being rewarded. As the student completes activities for the teacher and earns tokens, she will see that her card has some circles filled in while others are still empty. This arrangement helps give the student visual cues about how close she is to earning her reward and how much more work is still expected. Of course, it is important to keep in mind a student's age when setting goals for how long the student can be reasonably expected to work.

Putting It All Together

When each of these systems is introduced to a student, there are a number of important things for the teacher to arrange. Most important, a student using PECS must have continuous access to his own communication system. The communication book (or individual pictures) should never be hidden as a way of 'silencing' the student. That would be no different from putting tape over the mouth of a typically developing 2-year-old who asks too many questions. Next, a schedule system should be prepared to help the student navigate across the activities of the day. Finally, the student should be given a visual reinforcement system that will help in understanding what he is working for and how much more work is expected. By combining these systems, the student is given a solid basis by which he is better able to communicate with parents, teachers, and people in the community, and he is better able to understand his teachers' and parents' communicative efforts. Further, by stressing the visual aspects of these systems, the student has more continuous access to important information about what he can do and what is expected of him. A complete description of the elements involved in designing effective educational environments that incorporate PECS and other visual systems can be found in The Pyramid Approach to Education (Bondy, 1996).