Verbal Imitation

Verbal Imitation

The skills taught in the Verbal Imitation Program are important for all students with developmental delays, including students who do not currently speak and those who are echolalia. Through this program, the student learns to enunciate sounds, words, phrases, and sentences. Enunciation is the first step toward learning to use words in a meaningful manner, because once verbal imitation is mastered, this skill can be used to prompt responses in other programs that require verbal responses (e.g., Chapters 23 and 24). Verbal imitation training also aids the teacher in gaining control over the timing of an echolalia student's repetition of words and phrases. Whether the student is echolalia at the beginning of treatment or becomes echolalia as a result of the Verbal Imitation Program, echolalia responding may be so excessive that it interferes with the student's acquisition of vocal language. If the student you work with acquires this tendency, the interference of excessive echolalia responding may be reduced by using the techniques described in the section 'Managing Echolalia' toward the end of this chapter.

Gaining instructional control over the student's vocalizations accomplishes two goals. First, in shifting control over the student's vocalizations from the student's own sensory feedback (as in self-stimulatory behaviour) to the teacher's externally mediated reinforcers, the teacher is in a better position to subsequently influence and shape the student's vocalizations into recognizable and meaningful speech. In contrast, if the student's vocalizations are not effected by the teacher's use of extrinsic reinforcement, they are less likely to be brought under instructional control and less likely to be modified by the teacher. Second, once the student learns he can gain some control over his social environment by vocalizing, he makes a major step toward developing socially appropriate behaviours that may replace tantrums and self-stimulatory behaviours.

Given that parents and teachers tend to prioritize language skills, they may try to work quickly through the steps in this program. In teaching verbal imitation, the teachers need to constantly remind themselves that haste runs the risk of making language aversive to the student. This is a very difficult program to teach and a very difficult program for the student to master. Hence, everybody must be patient and move forward in small steps.

As a prerequisite for the current program, you should have established instructional control as described in Chapter 9. Specifically, the student should have learned to comply with elementary requests, such as sitting in a chair when asked to do so and should display few tantrums and self-stimulatory behaviours while in the teaching situation. The student should have also made progress in his mastery of matching and nonverbal imitation

(see Chapters 12 and 13) prior to beginning this program. Although there appears to be no discernible generalization between nonverbal imitation and verbal imitation, the prior learning of mouth movement imitations (e.g., closing lips, blowing, sticking out one's tongue) helps to prompt certain sounds. Finally, it is very important that you become proficient at employing discrimination learning procedures (Chapter 16), as these procedures are essential to teaching verbal imitation.

There are as of yet no data (e.g., regarding student characteristics) that allow teachers to predict whether the student is an auditory or a visual learner, a distinction that was described in Chapter 3. The only advice we have is to start teaching verbal imitation and see how well the student learns to imitate your vocalizations. If the student fails to learn to imitate speech or demonstrates extreme difficulty learning to do so, introduce Chapter 29 (Reading and Writing), Chapter 30 (Communication Strategies for Visual Learners), or both. Some preliminary and informal data suggest that some children learn to verbalize once progress is gained in the Reading and Writing Program.

The Verbal Imitation Program is divided into several phases, ranging from the relatively simple to the very complex, and can be outlined in brief as follows: In Phase 1, the student is taught to increase the amount of his vocalizations. This increase marks the first attempt to gain reinforcement control over the student's spontaneous and random vocalizations (e.g., babbling, humming) by shifting control from self-produced sensory reinforcers (as in self-stimulatory behaviour) to extrinsic reinforcers that you provide.

Phase 1 is the first step toward allowing you to influence and later shape the student's vocalizations into recognizable words.

Phase 2 builds upon Phase 1 and is a little more demanding in that the student does not obtain reinforcement simply for vocalizing but rather for vocalizing after you vocalize. That is, you make a sound, and if the student vocalizes shortly thereafter, the student is reinforced. In everyday language, the student is taught to listen and respond to your vocalization in order to be reinforced (in technical terms, the student is taught a temporal discrimination).

Phase 3 is complex because reinforcement in this phase is given only if the student's vocalizations match your vocalizations. For example, if you vocalize the sound 'ah,' the student is reinforced for imitating the sound 'ah.' If you then introduce another sound 'mm,' the student is provided with reinforcement contingent on his imitating the new sound 'mm.' These sounds (e.g., 'ah' and 'mm') are then subjected to discrimination learning procedures. This phase in particular marks the beginning of true verbal imitation and requires you to have thorough knowledge of discrimination learning principles.

Phase 4 goes beyond the imitation of isolated sounds and teaches the student to imitate a combination of sounds, such as those that compose simple words like 'dada' and 'baby.' Phase 5 teaches the imitation of complex words, and Phase 6 introduces strings of words as used in phrases and sentences (e.g., 'I want up'). Phase 7 teaches the student to imitate your use of volume, pitch, and speed.

Although we are reasonably confident in recommending that Phases 3, 4, 5, 6, and 7 be taught in consecutive order, there are no scientific data to indicate that Phase 1 must precede Phase 2 or that Phase 2 must precede Phase 3. Indeed, there are substantial individual differences among students. In our experience, some students have serious difficulties with Phase 1 but master Phase 2 relatively quickly. The sequence of phases in this chapter is the best guide we currently have to help you start verbal imitation training, aid in the student's success, and increase the student's motivation to talk. However, it is certainly not universally successful.

Because of the difficulty involved in teaching verbal imitation, we recommend that one or two teachers be appointed to specialize in this program. They should confer closely with one another and ask advice from the remaining team members and consultants. They should also inspect for inadvertent variations in teaching style across teachers, making sure that such variations do not occur and delay the student's mastery, especially in the early stages of verbal imitation training.

Phase 1: Increasing Vocalizations

To increase the frequency of the student's vocalizations, all you should do in Phase 1 is reinforce the student for vocalizing. A vocalization may be defined as any audible sound or word made by the student and includes grunting, laughing, coughing, babbling, 'ah,' 'ee,' 'baba,' recognizable words such as 'mama,' and the like. Within less than 1 second after the occurrence of a vocalization, a reinforcer should be delivered. Make sure that the rein-forcer is powerful and discriminable. For most students, we recommend food reinforcement (small bites of easily chewed food) and pronounced smiling, verbal approval, and clapping on the part of the teacher and other people present. In short, the student should cause quite an audience reaction based on her attempts to vocalize. Keep in mind, however, that some students are hypersensitive to sounds and may react with fear or anger to too much gusto, so set the volume accordingly.

If the student has a favourite food, it should be reserved for the Verbal Imitation Program. A small bite of a favourite food not only can serve to reinforce vocalizations but, as discussed later in this section, can prompt them as well, which paves the way for further opportunities to reinforce. The student's favourite food may make her content, and vocalizations are more likely to occur during periods of contentment. In short, the delivery of reinforcement (e.g., kisses, food, and tickles) often cues emotional behaviours, which in turn prompt more vocalizations. (Technically speaking, the reinforcing stimulus may possess unconditioned stimulus properties that may cue vocal respondent behaviours. These behaviours may later be brought under reinforcement control and become operant behaviour.)

Although the procedure in Phase 1 may seem relatively simple, you may encounter one or more of the following problems. First, the student's spontaneous rate of vocalizations may be quite low, giving you few opportunities to reinforce. Second, you may find it very difficult to prompt more frequent vocalizations. Third, given that vocalizations are fleeting responses, it may be difficult for the student to connect her response with the reinforcing stimulus. Technically speaking, it may be difficult for the student to discriminate the reinforcement contingency (the connection between the vocalization and the subsequent reinforcer), which makes it difficult for you to gain reinforcement control over the student's vocal behaviour.

Phase 1 should begin by having the student sit across from you, provided that such a situation does not make the student anxious. Whether and how long the student should remain seated should be contingent on her degree of comfort in such a situation. Perhaps 1- to 3-minute sessions interspersed with other programs and play may be optimal in the beginning.

There is no reason why the student should sit in the chair during Phase 1 except that such close proximity to the student gives you the opportunity to immediately reinforce each spontaneous vocalization. You may explore other physical arrangements that may be more conducive to eliciting vocalizations, such as having the student sit on your lap, on the floor or a couch with you, or on a swing or seesaw. Some students love taking baths and may vocalize readily there. Others may vocalize while watching videos or looking at books. All of these are examples of prompts.

Keep in mind that different students respond in varied ways to certain settings. For example, some students may be afraid of taking baths and stop vocalizing in such a setting. However, if you discover environments that promote vocalizations, there is no reason not to use such opportunities and later transfer gains back to the chair. In general, we recommend reinforcing spontaneous vocalizations throughout the day and then returning to the chair prior to beginning Phase 2.

It is often difficult to find prompts for vocal behaviour, and methods that work for one student may not work for another. It may be tempting to request vocalizations by saying such things as 'Talk,' 'Speak,' or the like. However, the student probably does not know the meaning of such requests at this stage into teaching. Therefore, these requests are likely to be ineffective and perhaps aversive because the student may have encountered such requests in the past and been frustrated by her failure to respond appropriately. Instead, use prompts such as saying, 'Hello,' while waving a hand and smiling, or holding a food reinforcer in front of the student while vocalizing (such as saying, 'Hello'), testing to see whether this gesture increases the probability of the student's vocalizing. If the student can imitate songs, she may be prompted to complete the words of a song after the teacher sings the first three or four words and then stops. If the student continues the song, such vocalizations should be reinforced.

Some prompts may be inherent in non-contingent use of positive reinforcers. Therefore, we advise that small bites of food be provided ('for free') every half minute or so during sessions of Phase 1 in hopes that these reinforcers cue contentment, and hence vocalizations, which should be reinforced with both social and food reinforcers. Also, try to prompt vocalizations by tickling the student, kissing the student on the neck, or stroking the student's cheeks or hair. Try activities such as helping the student jump up and down and turn upside down only if these are likely to make that particular student happy and more likely to vocalize (some students may become frightened). Smiles, cooing, gently poking the student's stomach, and otherwise giving signs of happiness may also be used as potential prompts. These behaviours may be similar to what adults employ in talking to very young typical children. The difference is that these behaviours are used in this case as prompts or as reinforcers that are given immediately after and contingent on the student's vocalizations.

Consult the student's family members for help in discovering effective prompts. Speech therapists also often possess useful information concerning methods for prompting speech. Some speech therapists are likely to recommend that the student be shown a favourite toy or similar object and then be required to vocalize in order to receive that toy. If done in a playful and non-demanding manner, this technique may prove useful.

For the skills learned in Phase 1 to be useful in subsequent phases, it seems reasonable to define mastery as two or more vocalizations per minute across 5 or more days in the setting where the training takes place (ideally this setting is the student sitting in the chair across from you at the table). We suggest working up to sessions of a 5- to 10-minute duration once an hour over the course of the day. There should be a noticeable increase in the student's rate of vocalizations over the level demonstrated prior to beginning Phase 1. Remember that the goal of Phase 1 is to increase the rate of vocalizations so that these vocalizations can be used in Phase 2.

If Phase 1 is successful, you have gained some control over vocalizations by the use of extrinsic reinforcers and are thereby in a better position to influence future language development. Keep in mind what was said earlier: It is difficult to gain reinforcement control over vocalizations because vocalizations are fleeting and thus the connection between the reinforcer and the behaviour may be difficult to make. Therefore, some students may not form the connection and hence may not make progress in Phase 1. Nevertheless, for reasons that are not understood at this time, some of the students who fail in Phase 1 succeed in Phase 2 or Phase 3. Whether or not the student makes notable progress in increasing her rate of vocalizations after a 2- to 3-week period, go on to Phase 2.

Phase 2: Bringing Vocalizations under Temporal Control

In the Early Receptive Language Program (Chapter 15), the student is taught to respond with nonverbal behaviours to your verbal requests. That is, the student learns to listen to you and behave as you ask him to behave. In Phase 2 of the Verbal Imitation Program, the student is taught to listen to your vocalizations and respond not with nonverbal behaviours but with vocalizations. This is similar to early baby talk when a parent talks to the new baby and pretends that the two are in a conversation.

Some readers may expect that if the student has made progress in the Early Receptive Language Program (i.e., the student has learned to respond to your vocal instructions by engaging in nonverbal responses), he should also have learned to respond to your vocalizations with vocal responses and should therefore master Phase 2 quickly. However, we have no clear evidence demonstrating that such a transfer occurs between these two programs. Listening to instructions and responding in the visual mode may not transfer to listening to instructions and responding in the auditory mode.

Phase 2 involves temporal discrimination learning in that you present the student with a vocal SD and reinforce any of the student's vocalizations that fall within a 5-second period following the SD. Any vocalizations that fall outside of this period during a sitting of Phase 2 should not be reinforced so that mastery of the temporal discrimination is facilitated. Technically speaking, your vocalization plus 5 seconds constitute the SD; the absence of this SD constitutes S Delta (the non-reinforced or negative stimulus). This procedure applies only to Phase 2 of the Verbal Imitation Program and does not mean that you should withhold reinforcement for the student's spontaneous vocalizations during other programs or free time.

Begin Phase 2 by sitting face to face with the student so the student can see your facial expressions. These visual cues may help the student make the discriminations, as well as prompt the student's vocalizations. Also, you should sit close to the student given that close proximity helps you provide immediate reinforcement for the student's correct responses.

► Step 1

Present the SD (e.g., 'Hello') and reinforce the student if he produces any vocal response within 5 seconds of the SD. Any and all of the student's vocalizations qualify as correct responses provided they fall within a 5-second period immediately following your vocal SD. If the student responds correctly, present the SD again about 2 to 3 seconds after the student receives the reinforcer from the preceding trial. If the student fails to respond, present the SD once every 2 to 3 seconds after the 5-second response interval. That way, the student will have approximately 20 opportunities to respond every minute. With such a high frequency of trials, it is likely that the student will learn to vocalize within the 5-second interval, especially if Phase 1 was mastered and the student's rate of vocalizations is high. If the student continues not to respond, pair the SD with a prompt as done in Phase 1 and fade the prompt over trials. Sittings of Phase 2 should be set at a similar duration to those of Phase 1. Intersperse other programs between sittings, including rehearsing and reinforcing already mastered programs so as to maintain the student's motivation. Mass trial the SD in Phase 2 until the student responds correctly in 9 out of 10 or 19 out of 20 unprompted trials.

► Step 2

One way to sharpen the temporal discrimination is by slowly decreasing the reinforcement contingency from responding within 5 seconds to responding within 4 seconds, then 3 seconds, and finally 2 seconds or less, provided mastery is met for each previous time interval. The reason for tightening the contingency to 1 or 2 seconds is that it gives clear evidence, for both you and the student, that mastery of the discrimination has been achieved.

We strongly recommend that you rehearse the mastery of Phase 2 for the next week or two in order to solidify the discrimination and help the student establish self-confidence and trust in you. During this time, the student's mastery should be generalized across teachers and environments.

Phase 3: Imitating Sounds

Phase 3 is designed to teach the student to imitate specific sounds. For example, if you say, 'ah,' the student should respond, 'ah'; likewise, if you say, 'mm,' the student should respond, 'mm.' This phase is important because the sounds that the student learns to imitate in this phase will be combined in subsequent phases to help the student imitate words and, later, phrases and sentences.

Deciding Which Sounds To Start With

There is typically some difficulty deciding with which sounds to begin teaching. We can offer three suggestions:

use sounds or words the student frequently vocalizes,

use sounds or words that can be prompted, and

Use sounds or words that are commonly overheard and that often occur early in the development of typical children.

Start by listing the sounds or words the student made in Phase 2 and those overheard to be made by her in her everyday environment. Some of these sounds may be simple, as in 'ooh' and 'ah,' or more complex, as in 'Winnie poo,' 'up,' 'no,' 'agogoo,' 'get out,' and 'dada.' Although the student is unlikely to imitate these sounds initially, you should begin with them, as the student may be more likely to imitate them than other sounds. The fact that these sounds are more likely to occur gives the student more opportunity for reinforcement and thus makes these sounds more likely to be imitated later on in this phase.

Most vocal sounds are difficult to prompt. Thus, teachers have fewer opportunities to reinforce and the student has fewer opportunities to learn in this program than in programs that do not require vocal responses. To complicate matters, the prompts used must facilitate one or more specific sounds. Because of the problem finding effective prompts, we discuss how to search for potential prompts.

Some students occasionally label letters or numbers printed on cards. Others complete words to parts of songs sung by their parents. Others may label items such as 'doggie,' 'cookie,' and so on, when these items appear as pictures in a book. In such instances, you may use the cards, songs, or pictures as prompts when presenting the SD, and later fade the printed material or songs to bring the student's vocalizations under the control of your vocalizations.

This strategy was effective for Robert, a boy diagnosed with autism. After several weeks of making no progress in making separate sounds, Robert's mother reported that he would spontaneously label letters shown in an alphabet book. The teachers used these letters (written on cards) as prompts for verbal imitation. The teachers gave the SD 'A' and simultaneously showed Robert a letter card with A written on it. With this prompt he was able to say, 'A.' The teachers then slowly faded out the word card until Robert imitated the teachers' verbal SD ('A') rather than responding to the visual prompt. A similar procedure was later used to prompt and teach the imitation of other letters.

If the student 'sings along' to certain songs, use the song as a prompt. For example, if a student already imitates a word like 'round' in the song 'The wheels on the bus go round and round,' present this song as the prompt and gradually reduce the number of words until only the word 'round' remains. 'Round' then become the SD. Reinforce any approximations of 'round' such as 'roond' or 'ound' and modify the SD to fit that approximation in the initial stages.

With many or most students, the teacher may fail to observe instances of spoken sounds or words whether they occur spontaneously or in response to specific stimuli such as songs, pictures, letters, and numbers. For such students, the sounds or parts of sounds may have to be manually prompted or prompted through nonverbal imitation. For example, a 'mm' sound may be prompted manually by the teacher gently closing the student's lips while presenting the SD 'mm' or by the teacher modelling closed lips (as taught in the Nonverbal Imitation Program, Chapter 13). Reinforcement should be provided while the student's lips are closed. Physical assistance and nonverbal imitation can also be used to occasion an open mouth, which may prompt an 'ah' sound. The prompt can then be faded by providing less and less physical assistance (for a manual prompt) or by performing less and less of the behaviour (for a modelling prompt). The teacher should use the partially correct responses of mouth closing (part of 'mm') and mouth opening (part of 'ah') as the starting point for shaping the sounds that correspond with these oral movements.

A relaxed and playful mood is likely to help prompt vocalizations, as was mentioned in the discussion of prompts for Phases 1 and 2. Remember that the student's contentment typically elicits vocal behaviour. In contrast, a student who is apprehensive and anxious may inhibit vocalizations (although these states may also increase instances of echolalia).

If you start with sounds that have visual components that can help prompt vocalizations (sounds such as 'ooh,' 'ah,' 'ee,' 'buh,' 'mm,' 'b,' and 'p'), it is likely that the student will lip read rather than respond to your sounds. Therefore, if such sounds are used, gradually fade the visual prompts until the sounds are presented when the student is not looking at you, your face is turned away from the student's line of vision, or your lips are covered with one hand or a piece of tissue. Remember to fade prompts gradually to maintain the student's correct responding.

Sounds that do not have clear visual components are difficult to prompt and may be more difficult for the student to discriminate than sounds that have visual components. However, once the student masters imitating vowels and consonants with visual components that can be partially prompted, the student may be more likely to master sounds (vowels and consonants) in which the visual components are harder to detect, as in 'g, 'c,' 's,' and 'k.' One way to prompt 'k' is by first teaching the student to imitate your cough, and then using differential reinforcement to separate out the 'k' sound from the cough sound. Labial-dental consonants such as 'd' and 't' are easier to teach because they can be prompted through the imitation of tongue and mouth positions. Although easier to teach separately, however, they are difficult to discriminate from one another. Note that '1,' 'r,' 'f,' and 'th' sounds are usually not correctly enunciated by typical children until quite a bit of language is gained. Therefore, postpone teaching these sounds until the student progresses further into this program. Even if the student does not master imitating all consonants, hundreds of useful words can be taught in this program (see Phase 5).

In determining which sounds to teach the student, consider the usual progression of typical children in addition to the suggestions made above. Early vowel sounds include 'aah,' 'ooh,' 'ee,' and 'uh.' Early consonant sounds include 'mm,' 'buh,' 'puh,' 'duh,' and 'tuh.' Intermediate vowel sounds include 'ay' (as in 'say') and 'i' (as in 'ice'). Intermediate consonant sounds include 'ss,' 'zz,' 'sh,' 'guh,' 'buh,' 'wuh,' 'yuh,' 'juh,' and 'ch.' More advanced sounds and discriminations include consonants such as 'fuh,' 'vuh,' 'lah,' 'ruh,' and 'th.'

Keep in mind that there are exceptions to the recommendations made in this section. Occasionally one encounters a student who quickly acquires mastery of guttural consonants such as 'k' and 'g' and slowly masters frontal vowels such as 'ah' and 'oh.' There are also some students who quickly acquire imitation of complex words or sound combinations (e.g., 'agogoo,' 'cookie,' 'heli-cop') even prior to their mastery of simple sounds.

Imitating Sounds

Arrange yourself and the student in a position in which the student was successful in Phase 2 of this program, preferably sitting across from each other in chairs. Remember, you want the student not only happy and relaxed, but also close enough to see your face in order to receive prompts as well as immediate reinforcement. Have your best reinforcers ready. It is a good idea to reinforce the student for coming to the chair and to let the student have one or two reinforcers for free while sitting nicely in the chair.

For illustrative purposes, we start with teaching the student to imitate the sounds 'ah' and 'mm.' 'Ah' and 'mm' sound different as well as look different when you present them (allowing visual cues to serve as prompts). However, 'ah' and 'mm' may not be ideal beginning sounds for the particular student you work with, and thus you should be flexible in your choice of early sounds to teach. For example, if the student regularly responds with 'eeh' to your 'ah,' then accept that response and change your SD1 to 'eeh.' If the student responds to 'mm' with 'beh,' then use 'beh' as SD2. The fact that both 'ah' and 'mm' (and 'eeh' and 'beh') sound and look different should facilitate the early discrimination. It is the discrimination between sounds that is key to successful imitation because it enables the student to attend to the similarity between her and your vocalizations. That is, the training helps the student match the two stimulus inputs, which requires that the student hear (discriminate, attend to) both her own and your vocalizations and learns to associate the two, a task that is not easily mastered and should not be taken for granted.

Because the first imitation tends to be particularly difficult for the student to master, leading to several non-reinforced trials and resulting frustration, occasionally intermix (e.g., every five to seven trials) mastered tasks that can be reinforced, such as imitating waving. Tantrums are not a bad sign unless the tantrums seriously interfere with the student's vocalizations or your attempts to prompt vocalizations. Make certain that you do not reinforce tantrums, as in letting the student's tantrum allow her to escape from the teaching situation. Verbal imitation is a difficult skill to acquire, and tantrums may indicate that the student is involved in the learning task.

In illustrating the following steps, SD1 is 'ah,' and the student's correct imitation of 'ah' is Rl. SD2 is 'mm,' and the student's correct imitation of 'mm' is R2. We advise you to avoid prefacing the SD with 'Say,' as in 'Say, 'ah,'' and, 'Say, 'mm.'' There are three reasons for this. First, the stimulus 'say' is common to both SDs and therefore serves to lessen the distinction between the SDs, making the discrimination more difficult for the student. Second, some students imitate 'say' along with the correct response (e.g., 'Say, 'ah''), and then 'say' must eventually be removed so that it does not interfere with the student's acquisition of appropriate imitations. Third, the student may not know the meaning of the word 'say.' Hence the word may be extraneous.

► Step 1

Present the SD1 ('ah') and mass trial Rl (the student's 'ah') using prompts, prompt fading, and differential reinforcement. Set mastery at 9 out of 10 or 19 out of 20 unprompted correct responses. To be reinforced, the student's Rl should sound close enough to your SD1 so that attending adults recognize it as an imitation, and it should occur shortly (i.e., within 5 seconds) after your SD1.

Step 2

Present the SD2 ('mm'), prompt, and mass trial R2 to mastery (9 out of 10 or 19 out of 20 unprompted correct responses). Anticipate that when switching to SD2 ('mm'), the student may respond with Rl ('ah') because this response was most recently reinforced. It is important that R2 ('mm') is a close approximation to your SD2 and maximally different from Rl ('ah'). If the sounds are not clearly distinguishable, take the time to shape their clarity rather than rushing to new sounds. If the vocalizations are not shaped to sound different from one another, they may combine into a sound such as 'mah.' If this occurs, the discrimination between Rl and R2 will be difficult in Step 3. Note that Step 3 requires firsthand and working familiarity with discrimination learning procedures (Chapter 16).

► Step 3

Intermix and differentially reinforce SD1-R1 and SD2-R2. If SD2-R2 ('mm') was presented and reinforced last, it is likely that the student will give R2 when you present SD1 ('ah'). To avoid this error, prompt and re-establish correct imitative responding to SD1 (3 unprompted correct imitations in a row), then switch to SD2, prompt, and re-establish R2 (3 unprompted correct imitations in row). Switch back to SD1 and re-establish Rl.

As you shift back and forth between SD1 and SD2 and employ differential reinforcement, the student gradually learns to discriminate between the two SDs as evidenced by fewer errors made and fewer prompts needed each time you shift from one SD to the other. In other words, with each successive contrast of SDs, the student becomes increasingly able to tell the sounds ('ah' and 'mm') apart through the reinforcement of SD1-R1 and SD2-R2 and the non-reinforcement of SD1-R2 and SD2-R1. As the strength of the correct associations increases, you can set mastery criterion at decreasing numbers of successive correct responses, such as moving from 3 in a row to 2 in a row and finally to 1, before switching SDs. This is done to avoid inadvertently reinforcing the student for repeating the same response (perseverating), a pattern of responding that interferes with the student's mastery of the contrast (discrimination) between SD1 and SD2. To ensure that the discrimination between SD1 and SD2 is accomplished, introduce random rotation after the student correctly responds to systematic alternation between SD1 and SD2 after single trials of each SD. Random rotation must be done to eliminate extraneous cues that may result in the student's learning a win-stay, lose-shift strategy rather than the discrimination.

Keep in mind that this first discrimination between sounds is the most difficult for the student to master. Therefore, once mastery of the discrimination between SD1 and SD2 is accomplished (9 out of 10 or 19 out of 20 unprompted correct responses with the SDs presented in random rotation), we recommend that this discrimination be practiced once every hour of formal teaching over the next 4 to 5 days, allowing the student access to a good deal of reinforcement. After this, generalize the discrimination across teachers and then situations. Teaching in different parts of the house, outside, and so forth, allows the gains to be spread to environments other than the original teaching setting. With each new situation (teacher or setting), a loss in mastery may occur, necessitating some prompting to re-establish mastery. To help reduce a loss in mastery, make sure that each teacher consistently adheres to the teaching procedures collectively agreed upon, and make certain that effective reinforcers are always employed. Continue generalization of this portion of the program for 2 to 3 weeks.

► Step 4

After 2 to 3 weeks of rehearsing the discrimination between SD1 and SD2 across teachers and environments and inter-mixing this discrimination with other tasks, introduce the third sound, SD3. SD3 should be as different from SD1 and SD2 as possible, while at the same time being a sound the student can master. As already stated, individual differences among students precludes our identifying a particular sound for all students. It may be safest to pick the third sound from the list of sounds made in Phases 1 and 2. 'Duh,' 'oh,' and 'puh' are examples of sounds that fulfil criteria of differing in both visual and auditory stimulus properties from the SDs used in our illustration of teaching the discrimination between 'ah' and 'mm.'

► Step 5 Onward

Teach SD3 as the first two SDs were taught. Once SD3 is mastered in mass trials, systematically intermix SD3 with SD1 and then with SD2 while maintaining the SD1-SD2 discrimination. Begin random rotation of the SDs once the student correctly responds to each SD with systematic alternation after single trials. Bring all three discriminations to mastery before SD4 and other new sounds are introduced.

Areas of Difficulty

If you use a visual prompting procedure (from the Nonverbal Imitation Program) by exaggerating mouth movements when vocalizing sounds, the student may imitate your movements but not the exact vocalizations by, for example, enunciating 'eh' rather than 'ah' in response to SD1. If this occurs, shift your SD1 to 'eh' in order to match the student's sound. The discrimination between SD1 and SD2 is important in this phase, not the actual sounds used. Make certain to remove (fade) the visual prompt.

It is difficult to specify when you should discontinue teaching a particular sound (e.g., SD1 'ah,' or some other sound) and move on to another sound. However, no progress toward mastery after 100 to 200 trials is typically a safe criterion. By no progress, we mean no approximations. You may return to the difficult sound at a later stage. If you fail to make progress with SD1 using visual prompting, discontinue that SD for the time being and go on to a sound that can be manually prompted. For example, switch to a sound like 'mm,' which you can manually prompt part of by holding the student's lips together when she vocalizes. You may also test for the student's ability to imitate other sounds by using different prompts.

If the student can imitate one sound (e.g., 'ah' or 'eh') but fails to make progress in imitating a second or third sound, shift to a contrasting stimulus such as blowing bubbles in imitation of your blowing bubbles. Fade the bubbles and use this remedial SD as the contrasting SD in discrimination learning procedures. As individual differences are enormous, it is a good idea to remain flexible in the sounds targeted and the means by which these sounds are prompted. For example, blowing bubbles did not serve as an effective prompt for a student who instead liked to blow out lit matches (perhaps the visual feedback from a flickering flame constituted a sensory reinforcer). The training steps proceeded as follows: (a) the teacher blew and concurrently presented the lit match for the student to blow out. The teacher reinforced the student's blowing with social approval, (b) the teacher continued to present trials but faded the prompt, presenting the match to the student without the flame. The student blew and was reinforced, (c) Over successive trials the teacher's hand was still presented, but the match was gradually faded behind the teacher's fingers. The student continued to blow and was reinforced. Finally, the teacher's hand was faded, bringing the student's blowing under the control of the teacher's SD. That is, the student learned to imitate the teacher's blowing and this response was then used as a contrasting stimulus in teaching imitations of sounds. Remember, it is the student's discriminations among the teacher's and her own vocalizations that are critical, and remedial (or pre-training) procedures may be helpful in the student's acquisition of these discriminations.

Shaping

As the student progresses in imitation training, it becomes increasingly important that the student's imitations match your vocalizations as closely as possible. It is highly likely that the accuracy of the imitation will vary over trials. Therefore, after the student shows a reliable but rough approximation of your SD, begin shaping that response to more closely match the SD. Specifically, on a given trial, reinforce the student's response only if it approximates your SD more closely than the last reinforced response. If the student fails to match the SD closely enough to be reinforced over 3 or 4 trials, backtrack and reinforce a less accurate approximation to keep the student's vocalization from being placed on extinction and to help prevent the student from losing interest in the imitation task. Also, if the student receives too little reinforcement, she is likely to tantrum. Be careful when providing informational 'No's' as consequences in this program because such consequences used early on in the Verbal Imitation Program may result in discouraging the student's vocalizations. If the student's approximations are not accurate enough to be positively reinforced, consider withholding reinforcement and repeating the SD.

Motivation

The Verbal Imitation Program may be stressful for the student. Therefore, intersperse SDs from mastered programs such as Matching, Nonverbal Imitation, or Receptive Language every 5 to 7 trials so as to amply reinforce the student for cooperating and to help ensure maintenance of the student's involvement. Try not to exceed five non-reinforced trials in a row before introducing a mastered task and keep the sessions short (they should not exceed 5 minutes). Remember to vary the reinforcers used within each sitting, especially if no one reinforcer is particularly strong. End each sitting with reinforcement for a correct response. Ending a sitting contingent on a correct response will provide two reinforcers: leaving a difficult situation and gaining access to a favourite edible, toy, or activity. Ending a sitting contingent on an incorrect response may serve to reinforce, and thus strengthen, incorrect responding.

Keeping Up a High Rate of Vocalizations

Remember to reinforce spontaneous vocalizations from Phases 1 and 2 to maintain the student's production of these sounds. Thus, should the student vocalize between sittings of Phase 3 of verbal imitation training or during other programs, reinforce such vocalizations by, for example, stating, 'Good talking,' while smiling. The best reinforcers, however, should be saved for correct imitations to keep the student motivated to perform the more difficult task of imitating rather than spontaneously vocalizing.

Loss of Mastery When Generalizing

Anticipate that some students may either stop responding or start making errors when a new teacher is introduced or when the teaching situation is changed to a different environment. If difficulties occur when a new teacher is introduced, gradually fade in the new teacher by having him or her sit beside the familiar teacher who presents the SDs. Next, with both teachers sitting next to one another, have the new teacher give the SDs at the same time as the familiar teacher. Then, if mastery is maintained, completely fade out the familiar teacher. If the relative abruptness of the last step results in incorrect responding by the student, fade the original teacher more gradually by having him or her move a foot or two away from the initial position after every few SDs.

To avoid deterioration in correct responding when attempts are made to generalize responding across new environments, change the environment in gradual, incremental steps. This can be done by moving away from the original arrangements of the chairs and table by first leaving the table but keeping the chairs, then moving to other chairs, then onto the floor, then to different parts of the original room, then to the adjoining hallway, and so on. Keep in mind that students with developmental delays are not the only ones who have difficulty generalizing; typically developing peoples evidence problems in generalizing behaviours across environments as well. Children do not always behave in public the way they behave at home. A way to guarantee discontinuity in any individual's behaviour is by quickly changing the persons and physical environments that surround her.

What to Expect

Prepare for the imitation of the first two sounds to be the most difficult, with a gradual increase in the rate of mastery over the next sounds introduced. For those students who enter the Verbal Imitation Program with previous (although perhaps occasional) expression of words or imitations of sounds or words, the rate of mastery tends to be more rapid than for those who have no history of vocalizations. If a student can express a word prior to imitation training, then that student already demonstrates evidence of being able to listen to and match verbal responses. There is currently no evidence that those students who imitated their parents' vocalizations and then stopped when they were approximately 18 to 24 months of age learn to master verbal imitation sooner than those who never imitated vocalizations.

One of the most promising signs occurs when, some time into verbal imitation training, the student begins to spontaneously imitate words she hears in everyday settings or the teacher's instructions in other programs that may run concurrently with the present program, such as 'sit down' or 'point to knee.' This gives evidence that the student is becoming echolalia and that the matching of the student's and the teacher's speech has acquired reinforcing properties for the student.

Note that some people become concerned about the emergence of echolalia speech and try to discourage it because it is considered a symptom of autism. It is a mistake to discourage or otherwise attempt to change echolalia until much later into treatment when the presence of echolalia is likely to hinder mainstreaming and other peer integration. It is likely that all persons echo the speech of others, but do so sub vocally (as in private speech), having been taught not to think out loud. For now, consider imitation as having become its own reinforcer; the student likes to talk and has become less dependent on the adult's use of extrinsic reinforcement.

Some students may be echolalia and thus already able to imitate words and strings of words upon entering treatment. However, such students may not echo (imitate) the teacher's words when asked to do so and may experience significant problems in learning to imitate the teacher's words when presented with these words in the Verbal Imitation Program. Perhaps spontaneous echolalia does not share the same properties as the vocal imitation taught in the present program.

An encouraging sign can be found in the Verbal Imitation Program, however, when examining the nature of the errors the student makes because these errors point to rational, learning-based origins. For example, suppose the student practices the three imitations 'ah,' 'mm,' and 'eeh.' Prior to mastery, the student 'mixes up' her answers such that if the teacher presents SD1 ('ah'), the student may not respond with Rl ('ah') but rather with R2 ('mm') or R3 ('eeh'). That is, the student makes the wrong associations. This occurs because SD1 ('ah') exercises some power (associative strength) over 'mm' or 'eeh.' This is known in the technical literature as stimulus generalization. In everyday language, it can be said that the student attempts, in a trial-and-error fashion, to hit upon the correct response. This should be considered a normal phenomenon. Through the continuation of discrimination training, the teacher withholds reinforcement for the incorrect associations (e.g., SD1-R3, SD3-R2) and thereby weakens them. At the same time, the teacher reinforces and strengthens the correct associations (SD1-R1, SD2-R2, and SD3-R3) to compete with the incorrect associations. Often, with each new imitation mastered, the number of total errors decreases until imitations of the teacher's vocalizations are made without error upon their respective first presentations. Technically speaking, this is referred to as generalized imitation.

Phase 4:

Imitation of Consonant-Vowel Combinations

Imitating the First Combination

Once the student learns to imitate his first 8 to 10 sounds, including vowels and some consonants, teach the student how to connect these sounds together into simple words or to combine consonants and vowels in ways that will later be needed to build words. Consonant-vowel combinations and simple words such as 'ma,' 'mu,' 'bah,' 'duh,' 'mama,' and 'dada' are examples of SDs you might start with. In the following steps, the procedures of imitation, shaping, and chaining are reviewed so as to help you use them interactively to build consonant-vowel combinations.

When to Teach and How to Progress

Reserve the majority of the verbal imitation sessions for the morning hours when most students' rates of learning appear to be maximized. Conduct sittings of an approximately 1- to 5-minute duration, with mastered tasks intermixed within the sitting. Give the student a 2- to 3-minute play break between sittings and intersperse sittings of programs that teach skills the student may more easily master than those involved in the Verbal Imitation Program (e.g., Nonverbal Imitation or Receptive Instructions).

Once the student can imitate between 5 and 10 sounds consistently, you may begin using an informational 'No' and withholding reinforcement if the student does not imitate the sounds you make or if the approximations the student gives are below the standard criterion the student previously attained. Do not make the 'No' loud and aversive. It should be given calmly and should not sound angry, frustrated, or disappointed, but rather should be given as informational feedback to let the student know that her response was incorrect and that she needs to try again after the next SD in order to obtain reinforcers. After the student masters the imitation of approximately 8 to 10 sounds (e.g., 5 vowels such as 'ooh,' 'eeh,' 'ah,' 'u,' 'oo,' and 2 to 3 consonants such as 'm,' 'd,' and 'b'), begin Phase 4, shaping the first simple syllables and words. At the same time, intersperse trials from Phase 3 to introduce new sounds. It is important to teach each new sound to mastery, one at a time, rather than introducing a number of sounds at once.

Shaping

A few students successfully imitate a consonantvowel combination the first time the teacher gives the SD. Other students make an approximation to the combination on the first few trials, and the teacher can then use shaping on the later trials to help the student more closely approximate the teacher's SD. Most likely there will be imperfect imitations during training, and you will have to use shaping procedures to help the student improve certain imitations. For example, you may express 'muh' to your 'mah' or express 'deh' to your 'dah.' To begin shaping, repeat the SD over several trials and reinforce the student's response on a given trial if it approximates your combination as closely as or more closely than the last reinforced response. It pays to shape in steps no larger than those that keep the student reinforced and successful. Prompt the student whenever possible, facilitating the approximation. Continue the shaping procedure until the student's approximation of the combination is consistently clear enough to be understood by other people. Remember that the student's response needs to be as close to your SD as possible. Otherwise, the student will not succeed in learning to match your SD or to discriminate between your SDs. Also remember that exposure to similarity among auditory stimuli (yours and the student's) will help establish similarity as a reinforcing event for the student and help develop generalized and spontaneous imitations.

Shaping procedures work only if there are some initial similarities between the student's and your vocalizations. In most instances, and especially with complex SDs, the student's response will not approximate your SD closely enough to allow for shaping. In such cases, your word or sound combination will have to be broken down into parts the student can imitate and then combine through chaining.

Chaining

In chaining, a teaching process introduced in Chapter 10, two or more responses are 'chained together' in such a fashion that when one response occurs, it cues a second response to follow. Technically speaking, one response generates stimuli for a second response. The chain can consist of simple units, as when 'm' provides the cue for 'a' to form the syllable 'ma.' In an example involving somewhat more complex stimuli, the response 'ma' may come to cue the student to repeat 'ma' to create the word 'mama.' Through the use of chaining procedures, separate words are later built into increasingly complex combinations such as phrases and sentences.

To chain sounds, a particular sound combination should be divided into component sounds that the student can already imitate, and then each sound should be presented as a separate trial. Thus, the combination 'mah' may be broken down into 'm' and 'ah.' The student should imitate each component when it is presented, and should be reinforced for repeating each component. Then slowly shorten the amount of time separating the presentation of each component and eventually provide reinforcement contingent on the student's completion of both components as one response.

In illustrating the following steps, SD1 and Rl represent 'm' and SD2 and R2 represent 'ah.' It is assumed that the student has already learned to correctly imitate your SD1 and SD2.

► Step 1

Present SD1 ('m') and reinforce the correct response.

► Step 2

Present SD2 ('ah') and reinforce the correct response.

► Step 3

Present SD1 and move to the next step (4) just as the student begins to respond.

► Step 4

Present SD2 and reinforce the correct response.

In Step 3 do not reinforce after the student's correct Rl but rather immediately present SD2. Withholding the reinforcer for Rl and immediately presenting reinforcement for correct responding to SD2 should help the combination 'm-ah' occur. Reinforcement can then be delivered contingent on gradually decreasing amounts of time between the two sounds ('m' and 'ah') until the student is able to respond correctly to the sounds given as a single SD ('mah').

The fact that the student learns to respond with 'mah' to your 'mah' does not mean that the student actually imitates your 'mah.' By contrasting 'mah' with another sound combination, you can determine whether the student is imitating the SD or merely memorizing a response. Again, discrimination learning is the key to your success as a teacher and to the student's mastery.

Imitating the Second Combination

The second consonant-vowel combination should be maximally different from the first combination in terms of both your sounds and your mouth movements. For example, if 'mah' is the first combination taught, 'beebee' is a good choice for the second combination. The second combination, like the first, should come from your list of frequently heard sounds and should be composed of sounds the student can imitate separately. This combination should be taught in the same way the first combination was taught. Be sure to continue rehearsing mastered combinations as each new combination is introduced, helping to maintain those responses while working on new combinations.

After mastery of the second combination presented alone, begin discrimination training with the first two combinations. Achieve mastery of SD1-R1 ('mah'), then proceed to Step 2 and achieve mastery of SD2-R2 ('beebee'). In Step 3, intermix SD1 and SD2 using differential reinforcement until the SDs are presented in random rotation. Continue to intermix the SDs until the student responds correctly in 9 out of 10 or 19 out of 20 unprompted trials. As done previously, solidify the discrimination among SD1 and SD2 by generalizing it across teachers and environments over the next 4 to 5 days.

Combinations 3 Through 8

Introduce the next six combinations (e.g., 'da' or 'dada,' 'up,' 'pa' or 'papa,' 'me') the same way the first and second combinations were introduced. Each time a new combination is acquired; this combination should be randomly rotated with all of the previously mastered combinations. For example, if the student mastered seven different combinations and his eighth combination is ready for random rotation, intermix the eighth combination with Combinations 1 through 4 the first time you randomize presentations, Combinations 5 through 7 the second time the SDs are randomized, and so on. This helps maintain the earlier mastered combinations or re-establish them.

Once the student masters six to eight combinations, move on to build words by following the procedures presented in Phase 5. The combinations the student practices in the current phase may already contain recognizable words (as in 'me') or they may be helpful in teaching words (e.g., by using 'mah' as a building block for teaching the word 'mama'). Although imitation of most words is more extensive than the combinations taught in this phase, the principles involved in teaching the student to imitate words are the same as those used to teach imitation of sounds and combinations.

To facilitate the student's acquisition of imitating words, it may be helpful to differentiate between homogeneous and heterogeneous chains. 'Mama,' 'papa,' and 'dada' are examples of homogeneous chains (the same combinations are repeated). 'Cookie,' 'table,' and 'baby' are examples of heterogeneous chains (different sounding components are combined). Homogeneous chains are likely to be easier to master; therefore, it is recommended that this type of combination be introduced first.

Note that when we refer to words, we do not imply that the student knows the meaning of the words he imitates. For example, if the student acquires imitation of the word 'mama,' he probably does not know what that word means. The meaning of words is taught separately in the receptive and expressive language programs.

Areas of Difficulty

If the student experiences difficulty acquiring a chain (e.g., the chain 'mah'), try the following variation. Present SD1 ('m'). After the student responds with 'm,' do not present SD2 ('ah') but rather pause and look expectantly at the student, waiting for R2 ('ah') to occur. The pause may operate as a prompt. If R2 occurs within 5 seconds of SD2, abundantly reinforce the student. The reason you present only 'm' instead of the combination 'mah' is that doing so increases the likelihood that the student will respond to the SD with 'm' and the pause with 'ah' and decreases the likelihood the student will respond to 'mah' as an SD with only 'ah.' Because 'ah' is the most recent sound the student hears when you say 'mah,' it is the most likely sound to be emitted. In other words, the 'ah' in 'mah' is likely to block the student's 'm.' The pause makes it more likely that the student will express both the 'm' and the 'ah,' saying, 'mah.'

The pause after 'm' may be considered a prompt for the student's 'ah.' If the pause (i.e., the time delay prompt) fails, add a visual prompt after 'm' by opening your mouth while looking expectantly at the student. Slowly fade out the visual prompt and shift to a pause prompt. Fade the prompt over subsequent trials until the student responds correctly in 9 out of 10 or 19 out of 20 unprompted trials. By the end of this step, your SD should be 'm' and a short pause. The student's correct response should be 'm-ah.'

Once the previous step is mastered, change the SD to 'mah.' To minimize errors, loudly present 'm' and then immediately present 'ah' in an almost inaudible volume. By presenting 'ah' quickly and at a low volume and not reinforcing the student for merely repeating 'ah,' the student is taught not to skip 'm' and merely repeat 'ah.' Abundantly reinforce the student for emitting 'm-ah' to your loud 'm' and barely audible 'ah.' Gradually increase the volume of 'ah' over subsequent trials until the student is eventually presented with the entire combination ('mah') stated at a constant volume. The speed with which 'ah' is increased to a normal volume is determined by the student; ideally it should be increased gradually enough that the student does not make errors but quickly enough that the student does not become prompt dependent.

Phase 5: Imitation of Words

The First Word

Use three criteria in selecting the initial word for the student to imitate. First, choose a word composed of sound combinations that will facilitate the student's mastery of imitation of the word. For instance, 'mama' or 'dada' should be chosen only if the student can imitate the component sounds of the word. Likewise, 'cookie' should be chosen only if the student can imitate 'koo' and 'key.' As mentioned earlier, because 'cookie' represents a heterogeneous chain, it may be more difficult to teach than 'dada' or 'mama, which is homogeneous chains. Second, choose a word that will sound different from the second word to be taught so as to facilitate the student's discrimination (and subsequent imitations). Third, whenever possible, select a word the student can use in her everyday environment to get things she wants. For example, 'open' can be used by the student to ask an adult to open a door. Similar functional use can be made of 'play' to leave the work table, 'mama' to go to her mother, 'up' to be lifted in the air, and the names of foods (e.g., cookie, bacon) to get her favourite snacks. It is not unusual to observe some students utter with great fidelity those words that lead to powerful reinforcers such as food or removal from the teaching situation.

The following words are examples of sound combinations that may be helpful to start with: 'mama,' 'papa' or 'dada,' 'bye-bye,' 'up,' 'baby,' 'tummy,' 'doggy,' 'cookie,' 'shoe,' 'nose,' and 'cup.' Note that words with sibilants (as in 'shoe'), guttural consonants (as in 'cookie'), and those requiring heterogeneous chains are relatively difficult to master.

Shaping

Start by exploring words or approximations of words the student has already mastered. For example, for the first few trials, give an SD (e.g., 'mama') and shape by reinforcing any approximations that include the main sounds of the word. Thus, 'ma,' 'mam,' 'am,' or 'mum' are adequate initial approximations of 'mama.' Likewise, an adequate initial approximation to 'dada' is 'daddy,' to 'doggy' is 'gogo,' and to 'baby' is 'baba.' To develop an approximation, you may want to break a word down into combinations the student has already mastered and practice these combinations separately. For example, if the student continues to imitate 'baba' when the teacher presents 'baby,' select out the 'bee' sound in 'baby' and practice it separately, later adding it back into the word 'baby.'

To shape words, follow the same procedures used to shape the combinations in Phase 4: Reinforce the student's response on a given trial only if it approximates your word as closely as or more closely than the last reinforced response. For example, if the target response is 'mama' and the student initially responds with 'mah,' reinforce her. If on the next trial the student responds only with 'm,' withhold reinforcement because 'm' is less close of an approximation than 'mah.' Once 'mah' is established, try withholding reinforcement to test for a closer approximation like 'mam,' 'mom,' or 'maha.' The withholding of the reinforcer is likely to occasion an increase in vocalizations. If a closer approximation occurs, reinforce the student. Withholding reinforcement for inadequate approximations can lead to the student's trying harder to come closer to the target response. If the student's performance deteriorates, go back and reinforce earlier sounds (like 'm') to reinstate the student's responding. It is almost inevitable that the student's performance will show ups and downs during shaping, but you must use reinforcement in a manner that ensures more ups than downs.

Continue the shaping procedure until the student's approximations of the words are consistently clear enough to be understood by most people. The student's responses need not be perfect at first, but they should be shaped to be as close to perfect as possible. For imitation training to achieve the maximum benefit, the similarity between the student's and your pronunciations must be very close. Otherwise, similarity will not become a reinforcing event for the student, and the student may then depend on extrinsic reinforcement such as food and praise to maintain language gains. If shaping fails to produce closer approximations of words, go on to chaining.

Chaining

The student may fail to approximate your SDs, providing insufficient opportunities for you to shape the student's responses. For example, the student may fail to imitate one or more of the component sounds that make up a word, inadequately approximating the word (e.g., the student may say, 'ah,' to 'papa' or 'doggy'). In such a case, the word must be built using a chaining procedure. Divide the word into its component sounds and present each sound as a separate trial as done with sound combinations. For example, 'papa' should be divided into 'p' and 'ah,' then 'pah,' then 'pah-pah,' and finally 'papa.' That is, chain 'p' and 'ah' together to form 'pah' (make SD1 'p' and SD2 'ah'). Then chain 'pah' and 'pah' together to form 'papa' (make SD1 'pah' and SD2 'pah'). It may be helpful to provide the teaching steps in detail even though there is considerable overlap from the steps described in Phase 4.

► Step 1

Present SD1 ('p') and reinforce the correct response.

EXPRESSIVE LANGUAGE

► Step 2

Present SD2 ('ah') and reinforce the correct response.

► Step 3

Present SD1 and move to the next step just as the student begins to respond.

► Step 4

Present SD2 and reinforce the correct response.

Repeat this sequence until virtually no pause prompt is left between the 'p' and the 'ah.' Once the student masters 'pah,' chain 'pah' and 'pah' together as follows:

► Step 5

Present SD1 ('pah') and reinforce the correct response.

► Step 6

Present SD2 ('pah') and reinforce the correct response.

► Step 7

Present SD1 and move to the next step just as the student begins to respond.

Step 8

Present SD2 and reinforce the correct response.

Step 9

Present SD3 ('papa'). If the student responds correctly, reinforce.

If the student fails to respond correctly, repeat Steps 7 and 8, gradually reducing the latency between the two 'pah' sounds until there is virtually no latency left between the first 'pah' and the second 'pah,' such that the student imitates 'papa.'

The Second Word

The second word chosen should be maximally different from the first word. For instance, if 'papa' is the first word taught, 'baby' may constitute an appropriate second word (even though it constitutes a heterogeneous chain). If possible, the second word, like the first, should come from your list of frequently heard words from the student or should be composed of sounds the student can imitate independently.

The second word should be taught in the same way the first word was taught. It may help to rehearse mastered words in separate sittings as each new word is introduced to help maintain the old responses and reduce confusion when working on new words. After the student correctly imitates the second word to mastery criterion, begin discrimination training with the first two words. That is, present SD1 (e.g., 'papa') and reinforce the student's Rl ('papa'). Prompt if necessary. Place mastery at 5 out of 5 or 9 out of 10 unprompted correct responses. Once mastery of SD1 is reached, present SD2 ('baby') and bring it to mastery. Next, intermix SD1 and SD2 using differential reinforcement and eventually place them into random rotation. Place mastery at 9 out of 10 or 19 out of 20 unprompted correct responses. At this point, generalize the discrimination across teachers and environments over the next 5 to 6 days.

Areas of Difficulty

Building the first word is a difficult procedure to write out in detail, and you may have to improvise a bit to help a particular student along. Some students may persist in making slight pauses between the component parts of a word for quite some time (e.g., 'pah' [pause] 'pah' rather than 'papa'). These pauses can be eliminated by the use of shaping such that you provide reinforcement for decreasing lengths of pauses.

Despite the most extensive efforts by well-trained teachers, many students fail to progress beyond imitation of separate sounds to imitation of words. In this regard, it is important to know that occasionally a student acquires vocal imitation of some words after mastering the matching of words as done in the Reading and Writing Program (Chapter 29). Similar observations have been reported by the use of the Picture Exchange Communication System (PECS) Program (Chapter 30). Whether these informal observations will hold up when subjected to an objective scientific inquiry remains to be seen. Nevertheless, it may be worthwhile to try either program to test for such advances.

Words 3 Through 8

Teach the next several words (e.g., 'mama,' 'doggy,' 'go,' 'cookie') the same way the first two were taught.

Once each new word is mastered, this word should be randomly rotated with all the previously mastered words. For example, if the student has seven different words mastered and her eighth word is ready for random rotation, we advise that you intermix this new word with Words 1 through 4 during the first randomization; with Words 5 through 7 during the second randomization; with Words 2, 3, 6, and 7 during the third randomization; and so on. Such intermixing helps the student to both discriminate new words from already established ones and rehearse and thereby maintain the mastery of established words.

So far we have limited the student's curriculum to short or relatively easy words. The teaching of more complex and polysyllabic words such as 'dinosaur,' 'umbrella,' 'caterpillar,' and 'helicopter' may be delayed for a month or more into training of this phase to avoid placing too much stress on the student. When such words are introduced, use shaping if the student accomplishes close approximations of the words. If the student does not produce close approximations, use chaining procedures as described previously.

Use the best reinforcers you have at your disposal for unprompted correct responses and extinction or informational 'No's' when the student is incorrect. It is tempting to move quickly once the student masters 8 to 10 word imitations given the importance of the mastery of verbal imitation. Although you receive a great deal of reinforcement for teaching efforts when the student progresses rapidly, it is important not to push the student too far too fast or to teach skills within this program at the expense of skills taught in other programs. We warn against moving too fast since the student cannot inform you, by language, that the load is too heavy. One risk losing progress gained because of the student's becoming noncompliant or non-responsive, tactics many students have mastered to perfection in the past.

Imitation of 8 to 10 words should be considered a major achievement and marks the occasion for you to increase the intermixing of other programs between sessions of the Verbal Imitation Program to expand the student's competence and reduce her stress level. Expressive language programs (Chapters 23 and 24) should be introduced to teach the student to use her newly acquired vocabulary to obtain a larger range of reinforcers in a more efficient manner than before (i.e., through the use of language; see Chapter 26 for methods of teaching the student to verbally indicate her desires). At the same time, gradually proceed through the Verbal Imitation Program, teaching new single sounds, combinations of sounds, and recognizable words to increase the student's vocabulary.

What to Expect

After the mastery of the first 10 to 20 words, many students begin to spontaneously echo the teacher's words. For example, in a receptive language program, the teacher may ask the student, 'Point to mommy,' and the student may echo some or all of the instruction. We refer to such students as auditory learners in contrast to visual learners. Auditory learners may be identified after 4 to 5 months of 40 hours per week of one-on-one intervention by a qualified teacher or after 9 to 12 months for a student who was less than 30 months old when the program began. Students who fail to master generalized verbal imitation (i.e., they do not become echolalia or become only partly so) vary widely in their progress in the Verbal Imitation Program. Some continue to acquire new words, although at a slow rate (e.g., they acquire one or two words every month). Others do not acquire any more words or acquire them sparsely and express these words only with considerable effort. The difficulties these students experience should be rightfully attributed to inadequacies in the Verbal Imitation Program rather than to hypothetical deficiencies of the student (e.g., apraxia, low muscle tone, central auditory processing deficits). The teaching of verbal imitation is a relatively new field; only since the mid-1960s has research been done on this very complex task.

Auditory learners proceed at a faster rate through the programs described in this manual. Most, but not all, of these students reach 'normal' functioning as defined by measures of treatment outcome (McEachin, Smith, & Lovaas, 1993), provided they receive 40 hours of one-on-one treatment by persons qualified to administer the UCLA treatment model.

Students who are primarily visual learners benefit mainly from visual forms of communication as described in the Reading and Writing Program (Chapter 29) or the PECS Program (Chapter 30). Nevertheless, we recommend continuing with the Verbal Imitation Program, although at a less intense rate. This is particularly important for students for whom the teacher discovers different and unique ways of teaching verbal imitation. For example, some students fail to make progress in verbal imitation as it is presented here, yet they imitate the voices of cartoon characters (e.g., from Disney's The Lion King). Many questions need to be answered to help isolate more effective ways of teaching verbal imitation than the methods presented in this chapter.

Even though the adaptation of visual forms of communication for students with developmental delays is in its infancy, you are likely to observe visual learners make extensive progress in communication through programs emphasizing visual stimuli. It should be noted that auditory learners also benefit from learning to read and write.

Phase 6: Imitation of Simple Phrases and Sentences

While teaching expressive language programs and maintaining and building upon earlier phases of the current program, teach (in separate sessions) a more complex imitative skill by introducing phrases and simple sentences. The student's imitation of the teacher's phrases and sentences can best be described as mastery of elaborate word chains or word strings and should not be understood to mean that the student has mastered sentences and phrases as grammatical units. Rather, the student's imitation of word strings can later be used to teach grammar. Procedures for teaching the student grammar and the use of sentence form to generate new and spontaneous sentences are presented in Chapter 26.

Phase 6 of the Verbal Imitation Program presents procedures for teaching the student to imitate phrases and sentences consisting of increasingly longer strings of words. Once the student acquires this skill, you can use such imitations to prompt correct answers in other programs and in everyday life. For example, when teaching an item in the Expressive Labelling of Objects Program (Chapter 23), the teacher may ask, 'What is it?' while pointing to a toy dog and prompt the student to answer, 'Brown dog,' through imitation. In a different program, the teacher may ask, 'What do you want?' and prompt the student to answer, through imitation of the teacher's vocalization, 'I want cookie.' To benefit from such prompts, the student must master imitation of the teacher's phrases and sentences.

Remember to teach the imitation of phrases and sentences that are functional for the student; that is, teach phrases and sentences that can be used by the student to secure immediate reinforcement. The following are examples of phrases and sentences that are helpful to start with. The words 'want' and 'juice' can be chained into the phrase 'Want juice.' The word 'I' can later be added to form the sentence 'I want juice.' Similarly, the words 'me,' 'big,' 'bird,' and 'give,' can be expanded to 'big bird' and later to 'Give me big bird.' The separate words 'I,' 'want,' 'play,' and 'to' can be expanded to the sentence 'I want play' and later to 'I want to play,' a sentence that the student can be taught to use to go outside and play.

Imitating the First Phrase and Sentence

Shaping

Introduce the first phrases and short sentences following the same basic procedure used to introduce the first word. Begin by presenting SD1, which may consist of the teacher expressing two-word combinations for the student to imitate. For example, give the SD 'Want eats' or any other two-word sequence for the student to approximate. Reinforce any approximations that include the main sounds or words of the phrase. You may want to explore ('probe') different short phrases and sentences and select the ones the student can match the closest. If the student can approximate imitating a three-word sentence (e.g., 'I want play'), select that sentence to teach first. Some two-word phrases may contain words that are more difficult for the student to approximate than simple three-word sentences. Start with combinations the student has the most success with and hence the most opportunity to be reinforced for and learn.

If the student makes adequate approximations of your SD on the first five or six presentations of that SD, use shaping on later trials to help the student more closely approximate your verbalizations. When shaping, reinforce the student's response on a given trial only if it approximates your phrase or sentence as closely as or more closely than the last reinforced response. If necessary, prompt by exaggerating sounds that are weak in the student's responses. If the student has problems with a three-word sentence (e.g., 'I want play'), simplify it to two words (e.g., 'Want play'). Continue the shaping procedure until the student's approximation of the two-word SD is consistently clear enough to be understood by most people, and then add the third word. If shaping does not produce closer approximations of your verbalizations, go on to chaining.

Chaining

For some phrases and sentences, the student may fail to even roughly approximate your SD on the first trials. For example, the student may fail to imitate more than one of the words that make up the phrase or sentence. If this occurs, you must build the phrase or sentence using one of two chaining procedures: forward chaining or backward chaining. As students differ in their responses to the two types of chaining procedures, deciding whether to use forward chaining or backward chaining requires a little experimentation. Try out both procedures to see which works best for the student you work with.

Forward Chaining.

Divide the phrase or sentence into its component words and present each word as a separate trial. In teaching a three-word chain, simply chain the first two words together and then add the third word. For example, the sentence 'I see ball' should first be divided into 'I,' then 'I see,' and finally 'I see ball.' Follow the same steps used to chain simple sounds into words. Chain together 'I' and 'see' (making SD1 'I' and SD2 'see'), then chain together 'I see' and 'ball' (making SD1 'I see' and SD2 'ball') as follows.

► Step 1

Present SD1 ('I') and reinforce the correct response.

► Step 2

Present SD2 ('see') and reinforce the correct response.

► Step 3

Present SD1 ('I') and go on to Step 4 just as the student begins to respond.

► Step 4

Present SD2 ('see') and reinforce the correct response.

► Step 5

Present 'I see,' and prompt by providing a short latency between the two words. Fade the prompt by gradually decreasing the latency between the two words until the phrase 'I see' is given as a single SD.

By following Steps 1 through 5, the student learns to express the response chain '1 see.' Once the student masters imitation of this phrase, chain 'I see' and 'ball' together to form a sentence.

► Step 6

Present SD1 ('1 see') and reinforce the correct response.

► Step 7

Present SD2 ('ball') reinforce the correct response.

► Step 8

Present SD1 and go on to the next step just as the student begins to respond.

► Step 9

Present SD2 and reinforce the correct response. Repeat Steps 9 and 10 until there is virtually no latency prompt left between 'I see' and 'ball' (e.g., 'I see [pause] ball,' then 'I see [less pause] ball,' and finally '1 see ball' [without a pause]).

Backward Chaining.

Backward chaining is beneficial for those students who tend to jump ahead, completing the response rather than imitating the teacher's sentence when forward chaining is used. For example, when using forward chaining to teach the sentence '1 see ball,' some students may say 'ball' when you ask them to imitate 'I see,' essentially completing the phrase you started. Backward chaining helps to eliminate this problem. To begin, chain together 'ball' and 'see' (make SD1 'ball' and SD2 'see'). Then chain together 'see ball' and 'I' (make SD1 'see ball' and SD2 'I'). The following is an example of backward chaining. Pay special attention to the order in which the SDs are presented.

► Step 1

Present SD1 ('ball') and reinforce the correct response.

► Step 2

Present SD2 ('see') and reinforce the correct response.

► Step 3

Present SD2 and move to the next step just as the student begins to respond.

► Step 4

Present SD1 and reinforce the correct response. Repeat Steps 3 and 4 until there is virtually no latency left between the presentations of 'see' and 'ball.' Once the phrase 'see ball' is mastered, chain 'I' and 'see ball' together.

► Step 5

Present SD1 ('see ball') and reinforce the correct response.

► Step 6

Present SD2 ('I') and reinforce the correct response.

► Step 7

Present SD2 and go on to the next step just as the student begins to respond.

► Step 8

Present SD1 and reinforce the correct response. Repeat Steps 7 and 8 until there is virtually no latency left between the word 'I' and the phrase 'see ball' (e.g., 'I [pause] see ball,' then 'I [less pause] see ball,' and finally 'I see ball' [without a pause]).

Imitating the Second Phrase and Sentence

The second phrase or sentence to be imitated should be maximally different from the first phrase or sentence. For instance, if 'I see ball' is taught first, 'Pick me up' is an appropriate second sentence to teach. The second phrase or sentence, like the first, should be composed of sounds and words the student can imitate on his own.

The second phrase or sentence should be taught in the same way the first phrase or sentence was taught. Be sure to continue working on the first phrase or sentence during separate sittings of verbal imitation as the second phrase or sentence is introduced. This helps maintain the mastered phrase or sentence and may reduce confusion when working on the new phrase or sentence. After the student correctly imitates the second phrase or sentence, begin discrimination training between the first and second phrases or sentences by following the discrimination training procedures described earlier.

Phrases and Sentences 3 Through 8

Teach the next six phrases or sentences (e.g., ' I love you,' 'I like juice,' 'I want cookie,' 'Long cold day,' 'Big red flower') the same way the first two were taught. As with the procedures used for teaching the imitation of words, you should include a number of mastered phrases and sentences in random rotation as each new phrase or sentence is ready to be randomly rotated. For example, if the student has seven different phrases or sentences mastered and the eighth is ready for random rotation, intermix this phrase or sentence with Phrases or Sentences 1 through 4 during the first randomization; with 5 through 7 during the second randomization; with 2, 3, 6, and 7 during the third randomization; and so on. While practicing skills from this phase of verbal imitation, remember to continue teaching new sounds and sound combinations from Phases 3 and 4 in separate sittings.

It is best to move rather slowly through the imitation of phrases and sentences. Try intermittently testing for generalized imitation by probing new word combinations after the student masters three-word combinations. Starting with a mean length of utterance (on average, the number of words the student can say in a chain) of three words, do not move on to four words until the student demonstrates mastery of 8 to 10 three-word chains across different settings and persons. Once the student gives evidence of imitating your presentation of novel three-word combinations, start teaching four-word combinations following the same procedures used to teach two- and three-word combinations. Use forward chaining (chaining Word 1 to Word 2 to Word 3 to Word 4) or backward chaining (chaining Word 4 to Word 3 to Word 2 to Word 1) or chain Word 1 to Word 2 and Word 3 to Word 4 separately, then chain the two groups of words together (chain the combination of Words 1 and 2 to the combination of Words 3 and 4).

Once the student masters imitating six to eight four-word phrases or sentences consistently, move on to Phase 7 to teach imitation of volume, pitch, and speed. Concurrently, the student should be taught to imitate new sounds, combinations, words, phrases, and sentences, including the names of people with whom the student interacts regularly and any other labels that are functional and useful for him in his everyday life.

Areas of Difficulty

The student may persist in making slight pauses between words in a sentence for quite some time (e.g., 'I see . . . ball'). These pauses should gradually diminish with increased training in other language programs. If they do not, they can be corrected by procedures outlined in Phase 7.

Some students may jump the gun by starting to imitate before you finish the sentence. If this occurs, it sometimes helps to stop the trial immediately and give an informational 'No' or to place your fingers on the student's lips until the SD is completed. If this fails, some sort of cue may be introduced as a prompt to help pace the student. For example, tapping out each word in a sentence with a block and then giving the block to the student to tap while completing his sentence is often an effective prompt. Another prompt can be provided by laying out three blocks in a row (if you are working with a three-word chain) and then pointing to one block with the expression of each word. For example, using this prompting procedure, if you say, 'I see ball,' you would touch the first block while saying, 'I,' the second block while saying, 'see,' and the third block while saying, 'ball.' The student is provided access to a similar row of blocks to help pace and prompt his sentence. Note that whether or not these prompts are effective, the student's learning to read sentences may help establish the concepts of pacing and chaining.

Do not add new phrases and sentences so quickly that the student has difficulty maintaining clear enunciation of each component word. Remember, for you to use the student's imitation of phrases and sentences as prompts in subsequent programs, the student's close match of your verbal prompts must be maintained. Two problems arise if this recommendation is ignored. First, you may find it necessary to correct the student's imitation of a particular word after he completes his sentence. As a result, the student's sentence is not reinforced and you may thus inadvertently place the student's imitation on extinction. If you need to sharpen the imitation of a particular word, take time to focus on this word in a separate session. Second, it is important to ensure a close imitation of the student's correct response in later programs so that no ambiguity in when or what to reinforce arises. For example, if the student is asked to label a picture 'Mommy and a baby,' and is prompted to do so but expresses the response as 'Me and baby,' the teacher will be uncertain as to whether to reinforce the response. Note that although typical children may be hard to understand when they are just beginning to talk, their behaviours do not extinguish as quickly as the appropriate behaviours gained by students with developmental delays. In short, it is in the student's best interest for you to secure adequate performance of early and basic steps rather than hurry forward to new and more complex programs.

Phase 7: Imitation of Volume, Pitch, and Speed

As the student makes progress in the earlier phases, problems may occur in the enunciation of volume, pitch, and overall speed of verbalizations. It is not unusual for some students, even after acquiring mastery in Phase 6, to express them very softly and almost inaudibly. Others may speak in a very monotonous voice, not changing expressions with variations in the content of sentences. Some students may space out their words, taking an unusually long time to complete a sentence. Some may hurry through a sentence so that it sounds like one word. Suggestions for overcoming these problems are presented in the sections that follow. Note that correction of these dimensions of speech should be introduced only after the student makes considerable progress both in verbal imitation and in expressive language programs (Chapters 23 and 24).

Imitation of Volume

To increase the volume of the student's responses, it may be helpful to teach her to discriminate between loud and soft voices. When teaching the student to imitate volume, use words or phrases that the student can already imitate and that are conducive to a loud or a soft voice. For example, 'roar' or 'yahoo' may both be good loud words, whereas 'baby' or 'kitty' may be good soft words. For the sake of illustration, let SD1 be 'Roar' stated in a loud voice and SD2 be 'Baby' stated in a whisper. Try prompting the word 'roar' to be said in a loud volume by actively gesturing like a cheerleader and by getting the student excited and moving around the room. Toy microphones also work well if you need a little help increasing the student's volume; many students are reinforced by hearing themselves speak into a microphone.

Step 1

Present SD1 ('Roar') and use shaping procedures to help the student match your loudness. That is, over successive trials, reinforce the student only if she imitates the word correctly and if the volume of her response is as loud as or louder than the volume of the last reinforced response. Set mastery at a criterion in which the student's loudness comes close to yours or distinctively above her earlier level of speaking.

► Step 2

Present SD2 ('Baby') at a very low decibel level, as in a whisper. Prompt the low volume by being very quiet and gentle, settling down yourself and the student. Put your hand or finger on the student's lips to help prompt a low volume. Reinforce the volume level if it was as soft as or softer than the loudness of the last reinforced response. Continue this procedure until the student closely matches your whisper. If the student does not exactly match your whisper, accept a difference in loudness between Rl and R2 that is pronounced enough to help the student to learn to discriminate between her own two levels of volume.

Step 3

Intermix SD1 and SD2 according to the discrimination learning paradigm. Set mastery at 9 out of 10 or 19 out of 20 unprompted correct responses with the SDs presented in random rotation. Remember to fade the physical prompts (e.g., gestures) for both volume levels.

Once imitation of each word with its respective volume level is achieved, you cannot guarantee that the student can discriminate loudness per se. Rather, the student may have reached mastery in Step 3 on the basis of responding to the two different words or some other cue. To ensure that the student can discriminate and imitate loudness, proceed as follows.

Step 4

Select a new word (e.g., 'hello') and present that word in a loud voice (as SD1), and then present the same word in a low voice (SD2). Prompt the initial trials, fade the prompts over successive trials, and then intermix the two SDs according to the discrimination learning paradigm. Finally, teach new words one at a time in a loud or a soft volume and then intermix the loudness of each word. For example, teach the word 'car' as a loud yell and then as a whisper, relying on discrimination learning procedures.

Bringing Volume under Instructional Control

Once the student learns to imitate the volume of the teacher's voice across many words, teachers, and environments, bring the loudness of the student's speech under instructional control. The goal of this step is for you to be able to tell the student to 'Speak louder' or 'Speak softer' as the situation demands (e.g., at home, in church, in restaurants). To accomplish this task, it is helpful to prompt by presenting the instructions in loud versus soft volumes. In brief, present the instruction 'Speak louder' with the word 'louder' at a high decibel level. Subsequently, the decibel level of your 'louder' is faded as a prompt while the student's loudness is maintained. The instruction 'Speak louder' is then contrasted with the instruction 'Speak softer' according to the following steps.

Step 1

Present SD1 ('Speak louder') and prompt the correct decibel level by stating 'louder' very loudly for the student to imitate (the student imitates the word 'louder' and its volume). When the student responds correctly in 5 out of 5 or 9 out of 10 trials, go on to the next step.

► Step 2

Present SD2 ('Speak softer') and prompt the student's correct response by stating the word 'softer' in a whisper. After the student reaches mastery criterion (5 out of 5 or 9 out of 10 correct responses), go on to the next step.

Step 3

Intermix SD1 and SD2 according to the discrimination learning paradigm. This step should proceed relatively quickly because the student already mastered loudness discrimination through the procedures in the previous section. Bring the discrimination among SD1 and SD2 to mastery at 9 of 10 or 19 out of 20 correct responses

Step 4

Gradually remove volume as a prompt, bringing the student's loudness under the control of your SD (i.e., 'Speak loud' or 'Speak soft' without any difference in volume among the two instructions). By maintaining differential reinforcement during prompt fading, the student is likely to learn to speak according to the instruction. That is, after you state, 'Speak loud,' in a conversational tone, the student should say, 'Loud,' loudly. Similarly, the student should respond with the appropriate decibel level when you instruct, 'Speak soft.' Set mastery at 5 out of 5 or 9 out of 10 unprompted correct responses for each instruction.

► Step 5

Generalize the instructions 'Speak loud' versus 'Speak soft' across new words and sentences. For example, if you instruct the student, 'Say loud hello,' the student should pronounce the word 'hello' loudly. To help the student refrain from imitating the 'say loud' component of the instruction, quickly and almost inaudibly state, 'Say loud,' while immediately prompting a loud 'hello' for the student to repeat. Gradually over the next several trials, fade in a normal volume of the words 'say loud' while maintaining, with differential reinforcement, the student's loud expression of the word 'hello.' Set mastery at 5 out of 5 or 9 out of 10 correct responses for each new instruction.

Step 6

Using the same techniques presented in Step 5, teach the student to express words in a low decibel level when you give new instructions such as 'Say soft bye-bye.' Set mastery at 5 out of 5 or 9 out of 10 correct responses for each new instruction.

► Step 7

Intermix two of the new SDs according to the discrimination learning paradigm. Set mastery at 9 out of 10 or 19 out of 20 unprompted correct responses. As with all programs that rely on discrimination learning, make certain that random rotation is included to avoid perseveration; win-stay, lose-shift; or other patterns of responding.

Over time, change the initial instructions to more typical expressions such as 'I can't hear you' or 'You are too loud' by slowly fading out the earlier instructions while gradually fading in the new instructions one word at a time.

Areas of Difficulty

It is difficult to maintain the student's volume level when volume is faded as a prompt (as in Step 4 of the previous section). This comes about because the student has been extensively reinforced for imitating several aspects of verbalizations, including volume. Nevertheless, with patience and practice, the student should master this program. If the student has extreme difficulty with volume training, however, hold off on teaching this skill unless the student speaks so softly that she cannot easily be heard by the teacher.

Imitation of Pitch and Contour

The portion of the program for teaching the student to modulate pitch and contour in an appropriate fashion is also taught through imitation and may be appropriate to introduce here, even though the implementation of this program should be delayed until substantial progress in expressive language is demonstrated. Pitch refers to the frequency with which sound waves are produced and is analogous to what are called deep versus high voices (as in a bass vs. a soprano voice). Contour refers to the modulation of pitch across parts or types of sentences (such as a question, which ends in a rising pitch). The speech of persons with autism is often characterized as monotone and lacking in contour or expression; the teaching of pitch and contour modulation helps eliminate this characteristic.

The procedures for teaching appropriate variation in pitch are similar to the procedures used for teaching loud-ness imitation. When teaching the student to imitate pitch, use words or phrases the student can already imitate and those that may help prompt a high or a low pitch. For example, 'hi,' 'wee,' or 'cook' may help generate high pitches, whereas 'cow,' 'no,' or 'daddy' may facilitate low pitches. You may also attempt to prompt a high pitch by sitting up tall and raising your arms while smiling. You may prompt a lower pitch by slouching or changing your facial expression to a frown. As with all vocal imitations, you may find it difficult to isolate effective prompts and may have to rely heavily on shaping. To illustrate the steps that follow, let SD1 be 'Wee' stated in a high-pitched voice and let SD2 be 'Daddy' stated in a low-pitched voice.

Step 1

Present SD1 ('Wee') in a high-pitched voice. Over a series of trials, shape the student's response to match your high pitch by reinforcing closer and closer approximations of your pitch.

Step 2

Present SD2 ('Daddy') in a low-pitched voice. Shape the student's pitch to resemble your pitch. It is unlikely that the student will achieve a perfect match of your pitch in Steps 1 and 2. Perfection in pitch imitation is not necessary, however, as long as the adults who are present agree that the student's pitch of SD1 is noticeably different from that of SD2. By noticeably different, we mean that the difference is large enough to enable the student to make the discrimination in Step 3.

Step 3

Intermix SD1 and SD2 according to the discrimination learning paradigm. Place mastery at 9 out of 10 or 19 out of 20 unprompted correct responses.

► Step 4

Begin generalization training across words. That is, replace 'Wee' in SD1 and 'Daddy' in SD2 with new words and train to criterion as done in Steps 1 through 3. Later, present the same word as both SD1 and SD2 but change the pitch in which the word is stated. Both procedures help generalize pitch across words by separating particular pitches from particular words.

► Step 5

After the student imitates the pitch of words the first time they are given, present a three-sound SD1 such as 'da-dee-da,' expressing the sounds at different pitches (e.g., low-high-low). Shape the student's response until she imitates the pattern of pitches you make. When the student masters SD1, present SD2 with a different pattern of pitches and shape the student's pitches to match yours. Next, introduce new three-part sounds and shape the student's responses until the student imitates patterns of pitches the first time they are presented. The final step is to take one three-part SD and change it randomly. For example, take SD1 ('da-dee-da') and change the pitch of that SD (e.g., high-high-low, then high-low-high, then low-low-high, etc.).

Use these procedures to teach the student to imitate questions ('What is it?' 'What did you say?' and 'How are you?'), which require a rising pitch across words. In the same manner, the student may be taught to express declarative statements (e.g., 'I don't want that' and 'My name is Ben') with a falling pitch.

Imitation of Speed

Some students speak very slowly and do not adjust the speed of their expression to contexts. As with volume and pitch, speed of utterances may have to be taught separately as follows.

► Step 1

Present SD1, which consists of any mastered three-syllable word or a word sequence expressed at a rapid rate. An example is 'da-da-da' stated so that it consumes approximately 1 second. Shape the student's speed of expression to closely resemble that of your speed. Speed may be prompted by the use of visual cues such as gesturing in rhythm to your speech. Set mastery at 5 out of 5 or 9 out of 10 unprompted correct responses.

► Step 2

Present SD2, a sequence of syllables or words similar in length to SD1 but different in sound (e.g., 'tee-tee-tee'). Say SD2 at a rate that lasts approximately 3 seconds. Maximize the difference between SD1 and SD2 by exaggerating the slow versus fast speed. Prompt and shape the student's slow rate to approximate yours. Set mastery at 5 out of 5 or 9 out of 10 unprompted correct responses.

► Step 3

Intermix SD1 and SD2 and bring to mastery (9 out of 10 or 19 out of 20 unprompted correct responses), adhering to the discrimination paradigm.

In subsequent steps, generalize speed to different syllable or word combinations. Once this is mastered, bring the student's rate of speech under your instructional control by teaching her to respond correctly to the instructions 'Speak faster' versus 'Speak slower.' This may be accomplished by following procedures similar to those presented under the section 'Bringing Volume under Instructional Control.'

Areas of Difficulty

Be aware that the student may regress during the Verbal Imitation Program. The student may start to tantrum and self-stimulate more than he did before this program was started, or the student may achieve a level of mastery one week that may not be improved upon the following week, or he may lose some of the previous week's gains. Do not be alarmed by such occurrences. Verbal Imitation is one of the most elaborate programs to teach and one of the most difficult programs for the student to learn. Hence, this program can be very frustrating for the both you and the student. We know of no program in which the student continuously improves from session to session without occasional setbacks. If regression does occur, simply ease back into the program and return to a stage the student can master and thus be reinforced for so as to regain his confidence and feeling of success. You may also want to consider cutting in half the length of time you keep the student in the chair. Intersperse various instructions from other programs the student can successfully perform to help him regain the feeling of success and break up the monotony brought on by the repetition of a single program. You may also try changing reinforcers or teaching in a different room of the house.

As the student's imitative speech is subjected to programs involving grammar and meaning, it is likely that the clarity of his enunciation will suffer. It may become increasingly difficult to understand the student and decide whether to reinforce an utterance. In such cases, one must go back to the Verbal Imitation Program to reinstate adequate enunciation.

There are several ways to help the student maintain and achieve well-enunciated speech. One such procedure consists of teaching the student to read (see Chapter 29). The visual cues involved in reading may help improve the student's enunciation. Teaching the student to be the teacher (while you play the student) is another way to maintain and achieve well-enunciated speech; it is quite common to observe a student give instructions to the attending teacher in a most precise manner and with adequate volume. Perhaps this is because the student is reinforced by taking control.

Managing Echolalia

Imitation is a basic mechanism by which to learn. Typical students often learn the correct answers to questions by listening to and reproducing the teacher's answers or other students' answers. For example, if a teacher asks a typical student, 'When is Washington's birthday?' and the student fails to answer, either the teacher or other students in the class will prompt the answer, 'February 22.' The teacher may then again ask the question to the student who failed to answer and teach the student to repeat the answer ('February 22'). In contrast, an echolalia student will repeat the question, 'When is Washington's birthday?' and fail to associate the correct answer. For example, suppose the teacher asks the student, 'What is your name?' and the student echoes the entire question or only a portion of the question, such as 'Name' or 'Your name.' The teacher may try to prompt the correct answer by stating the student's name (e.g., 'Michael') in hopes that when the question is repeated, the student will answer correctly. In all likelihood, the student will again repeat the teacher's question and not the correct answer. The student's repetition of the teacher's question serves to block or prevent learning of the correct answer.

The following procedures are recommended for helping the student inhibit inappropriate echolalia responding. One procedure consists of using a volume prompt as a way to help the student discriminate between what to repeat and what not to repeat. For example, if the teacher asks, 'What is your name?' and wants to prompt the student to answer appropriately (e.g., 'Michael'), the teacher may quickly whisper the question, 'What is your name?' and then immediately and loudly prompt 'Michael.' Directly after the student says, 'Michael,' the teacher should provide the student with abundant reinforcement. In subsequent trials, the teacher should fade the prompt by slowly raising the volume and decreasing the speed of the question while slowly lessening the volume of the correct answer.

Generalize the student's imitation of correct answers by introducing new questions that must be prompted in order to be answered correctly. For this procedure to be effective, you must present numerous short, simple questions that require one-word answers. The loudness of the correct answer to each question should be faded over trials. Contemporaneous with the presentation of novel questions, the student is likely to generalize imitation of the correct responses rather than merely repeat your questions. Should the student fail, return to earlier steps by lowering the volume of the question while raising the volume of the answer. Reinstate the prompt fading process once correct responding is established.

A second procedure consists of teaching the student to answer 'I don't know' in response to questions that are unfamiliar to her. To illustrate, the teacher may ask, 'What is the capital of France?' and then immediately prompt the student to repeat 'I don't know.' Present the question in a low volume and the answer in a high volume, and immediately reinforce the student for repeating the correct answer (i.e., 'I don't know'). In gradual steps, increase the volume of the question and decrease the volume of the prompted answer. Repeat this question-and-answer sequence until the student answers correctly five times in a row without your prompt. Then go on to additional questions, such as 'Where does dad work?' 'Why do we have fire-fighters?' and 'Why do we eat food?' Once the student's responses to several such SDs are mastered, help the student discriminate between what she knows and does not know by intermixing questions she can already answer. Adhere to discrimination learning procedures when teaching this skill.

Further aid the student by teaching her to say, 'What is the answer?' 'Help me,' or 'Tell me' in response to unknown SDs. One way to accomplish this is by prompting the statement 'Tell me' after the student responds with 'I don't know' to a question such as 'Why do we eat food?' This sequence should be repeated until the student can answer with 'Tell me' without being prompted. The completed, mastered sequence would proceed as follows:

TEACHER: Why do we eat? STUDENT: I don't know, tell me. TEACHER: To live.

Over successive trials, the student is provided with the opportunity to generalize this kind of verbal interchange. The rate of acquisition varies across students; some students master and generalize this task after three or four questions, whereas others show little or no progress.

Concluding Comments

The procedures described in this chapter are complex, may seem tedious to apply, and may seem to achieve limited gains. It is important to remember, however, that once the student masters most of the beginning steps in this program, it takes relatively less work to expand the student's mastery across new and advanced programs. Furthermore, once gains are made, teaching can be taken out of the highly structured setting necessary for mastery of the first discriminations, and new skills can be taught and expanded in less formal environments. As the student's imitations become more generalized and inherently reinforcing, you should observe the student learning from the volume, inflection, and speed expressed by adults and typical peers in his everyday environment.

Keep in mind that the student cannot maintain and generalize what he learns in the Verbal Imitation Program unless these skills are practiced outside of scheduled teaching sessions. Therefore, provide many informal rehearsals of the skills taught to the student through this program. Present instructions in a playful and enjoyable context, increasing the likelihood that the student will learn to like to imitate speech. If the student learns to enjoy imitating speech, he may become echolalia, which is good in the early stages of teaching. The student may then 'play' with speech, even if he does not know the meaning of what he says. Meaning can be taught later. There are reasons to believe that students who echo the teacher's instructions learn how to respond to such instructions sooner than students who do not echo. Apparently, the student who echoes is able to 'store' the teacher's instructions and thus is better able to respond to those instructions; the instructions are not gone the moment the teacher finishes them. It may be helpful to consider that we all tend to echo questions to which we do not know the answer, or to questions that are complex with answers not immediately available. We often echo these questions sub vocally, rehearsing them to give us time to formulate the correct answer.

The Verbal Imitation Program tends to remain a focus of teaching for a relatively long time as this program helps increase vocalizations and fix enunciation problems. Verbal imitation is also used to prompt answers to questions throughout much of the student's early years. It is therefore important for both the student and the teacher to become proficient at this program.

The teaching skills learned in the current program help the teaching of most, if not all, of the other programs presented in this manual. With mastery of the procedures presented in this program, the teacher takes a giant step toward becoming an expert at teaching the student with developmental delays.