Beginning Language Training (Mand Training)



Chapter 9

Beginning Mand Training

(DRAFT VERSION)

April 13, 2005

The Importance of the Mand

The first type of verbal behavior emitted by an infant usually occurs between 4 and 8 weeks of life (Bijou & Baer, 1965; Schlinger, 1998; Skinner, 1957). Linguistic behavior has its origins in the form of unlearned crying which is a reflex related to biological needs, but soon crying becomes a learned behavior as it gets reinforced each day (Skinner, 1957). For example, infants cry when they are hungry and as a result receive food. They also cry when they are uncomfortable or tired and receive comfort, they cry when they are cold and receive warmth, they cry when they are in pain or afraid and receive relief, and so on. Initially, most parents play a guessing game to determine why their child is crying. But, eventually different forms of crying begin to emerge for each type of motivation and parents are able to identify a specific cry as related to hunger, pain, sleepiness, etc. Thus, early language development is a product of the interaction between biological and environmental variables. Around 8 weeks of life infants even develop what has been identified in the literature as “fake crying” (Novak, 1996; Wolff, 1969). These cries are not related to biological needs, but are requests for adult attention. For this type of cry, the motivation is related to wanting adult attention and the consequences consist of obtaining that attention. For most infants, crying becomes a way to communicate with adults in order to obtain desired reinforcers and to remove aversive stimuli.

Skinner (1957) identifies this type of verbal behavior as a mand. A mand is a type of language that involves a three-component interaction between a speaker and a listener. First, there must be some kind of motivation affecting the speaker (e.g., the child is hungry), second, this motivation evokes some form of behavior (e.g., crying), and third, this behavior is followed by a consequence delivered by a listener that is directly related to the motivation (e.g., food). In behavioral psychology the relation between antecedents, behavior, and consequences is called the three-term contingency (Skinner, 1953). Each time this three-term relation occurs the behavior in the middle may change slightly. The crying may get louder, sharper, higher pitched, etc., depending on which aspect of the cry is most consistently followed by the delivery of food. This process is called shaping. Behavior is constantly shaped by the effects that it has on the social environment. The behavior of a child who is hungry is gradually being shaped by the consequences delivered by adults in his environment. The child learns to emit a specific form of crying for hunger that is different from the cry that is related to fear, or the cry that is related to being sleepy, because of differential consequences. That is, the reinforcer related to food deprivation is food, while the reinforcer related to sleep deprivation might be cuddling and rocking.

Early infant manding may soon begin to occur with a variety of response forms such as single phonemes, intonations, grunting, grimacing, as well as negative behaviors such as screaming, crying and tantruming. Infants typically emit thousands of mands during their first year of life. This high rate of manding plays a key role in establishing the foundation for further language development. Not only does manding get a child what he wants, but also the roles of speaker and listener become firmly established. An infant learns that manding, or “speaking” brings about changes in adult behavior and allows the infant to control the social environment. In other words, with manding the infant is an active participant in day-to-day events, and has a say in those events. The infant, of course, does not identify the fact that he can control adult behavior, but this effect becomes quite obvious to the parent, especially those whose infant’s rate of manding becomes excessive, as may be observed with infants identified as “colic.”

Manding serves several other important functions for a child and parent (Table 5-1). In addition to getting a child what he wants, much of what is typically identified as bonding involves the association of adults with the presentation of reinforcers, and the removal of aversive stimuli. Successful manding allows a child to inform the parent of important events that are currently affecting a child such as hunger, pain, and fearful events. Parents are then able to provide the appropriate consequence at a point in time when that specific consequence is most effective, thus increasing the bonding process. For example, if a child is frightened by an approaching dog and emits a cry that is identified by the parents as a mand for help, the parent can immediately remove the aversive stimulus by picking up the child. An important effect of this interaction is that the parent is paired with the removal of an aversive event for the child, which increases a parent’s conditioned reinforcement value.

Eventually, around 8 to 12 months of life an infant begins to emit approximations to understandable words as mands. When words become more clearly understood by listeners, specific consequences are delivered more consistently and the frequency and variation of mands begins to increase substantially. Many of the mands are related to the delivery of reinforcers (e.g., Juice, Cookie. Elmo), but perhaps an equal amount of mands are related to the removal of aversives (No, Head shakes, Mine, Screams, Tantrums, etc.). Skinner (1957) points out that the mand is the only type of verbal behavior that directly benefits the speaker. This is because the mand is the only type of verbal behavior that is evoked by the speaker’s motivation and has consequences that are specific to that motivation. Manding allows a child to control his social environment to some degree, and bring about changes in that environment. The other types of verbal behavior (e.g., echoic, tact, intraverbal) are for the benefit of the listener. Even when a child begins to acquire the other types of verbal behavior, manding still dominates as the most frequent verbal operant. Most parents have no trouble in agreeing that day-to-day life with a 12-18 month old infant is busy, and it is typically all about what the child wants to do.

Table 1

-------------------------------------------------------------------------------------------------------------------

Importance of the Mand

Mands are the first type of verbal behavior acquired by typical children

Manding is the only verbal operant that directly benefits the speaker

Manding allows a child to get what he wants, when it is wanted

A parent or caretaker is paired with the delivery of reinforcement related to a specific mand

Manding brings about desired changes or conditions

Manding allows a child to control the social environment

Manding training can decrease negative behaviors that serve the mand function

Mand training helps to establish speaker as well as listener rolls

Mand trials can be used as reinforcers for other forms of verbal behavior

Manding is essential for social interaction

Manding allows a speaker to acquire new information and new forms of verbal behavior

Neglect of the mand can impair language development

Mands do not emerge by training on the other verbal operants

Manding allows one to get rid of what he/she does NOT want, when it is not wanted

Neglect of the mand can result in emotional impairment

Excessive manding is a burden on the listener

-------------------------------------------------------------------------------------------------------------------

Neglect or Incorrect Reinforcement of the Mand

Neglecting or incorrectly reinforcing a child’s mand repertoire can have serious developmental consequences (Drash & Tutor, 1993; 2004; Skinner, 1957). There are a variety of potential problems that can lead to linguistic, behavioral, and emotional impairments for a child. The effects of failure to reinforce mands that are related to biological needs are clear, and have implications for infant survival. But as an infant grows there are several potential problems that may be related to failing to reinforce mands or delivering incorrect consequence related to an infant’s mand repertoire. Not only do parents lose out on important pairing opportunities, but if acceptable responses forms are not established for manding (i.e., words), it is common to observe negative behavior that functions as a mand. These problems are often identified with labels such as a child who is a brat, crybaby, or colicky. Sometimes the negative behavior can become so severe (e.g., tantrums, aggression, self-injurious behavior) that parents are at a complete loss as to what to do. It is also possible that the failure to correctly reinforce mands may produce a reduction in attempts to mand (extinction), and a child may seem indifferent, withdrawn, or asocial. Or, the child may develop a very restricted range of mands that may only occur when the MO is quite strong. On the other hand, excessive reinforcement of mands may produce a mand rate that becomes intolerable to parents and caretakers, and these children may be identified as “spoiled” or “needy.” Therefore, it is very important for parents to understand the role of mands for an infant, and to provide a balanced environment between the three extremes of ignoring or punishing mands, reinforcing negative mands, or over-reinforcement of mands.

Where to Start Mand Training for a Nonverbal Child

Given the importance of manding to typical language and social development, and the strong establishment of manding prior to the other forms of verbal behavior, it seems quite reasonable that language intervention should begin with mand training for a child who is nonverbal. The other early skills (vocal and motor imitation, receptive language, matching, and naming) are certainly important, but clearly not as important as a functioning mand repertoire for a nonverbal child. If a child has some of these other repertoires, these skills can facilitate mand acquisition (e.g., echoic and motor imitation), but without direct training it is unlikely that manding will emerge from these other repertoires (Hall & Sundberg, 1987; Twyman, 1992). Training designed to advance these other skills can occur concurrently with mand training, if a child can already perform them (see below).

The most critical aspect of mand training is the fact that the response needs to be under the functional control of what is currently motivating to the child. Therefore, mand training can only occur when the relevant motivating operation (MO) is strong, that is, when the child actually wants the targeted item or activity. For example, if a child demonstrates an MO for watching bubbles by reaching for the bubbles, then smiling and laughing as he pops the bubbles in the air, the timing is probably right to teach him a mand for bubbles. When the child is no longer interested in bubbles (motivation is lost) mand training for bubbles is over, and the search for new motivators begins (see below for more detail on adding new words). Another complication of mand training is that different words (i.e., different response forms) need to be established and brought under the control of different MO. Vocal words are of course the most common response form, but sign language, pictures, or written words can also be used. The current chapter will focus on the use of speech as a response form, while the following chapter will describe procedures for using the alternatives to speech.

Selecting the First Words

The first words that are chosen for mand training need to be carefully selected. First, there are several words that should NOT be selected as the first words to teach (Table 5-2). Under no circumstances, should the first mand targeted consist of words such as More, Mine, Please, Yes, or Pointing. These responses may be acquired quickly because they are followed by a wide variety of reinforcers and become a generalized mand. The response becomes equal to I want something, but if the item is not present (e.g., the child is in a car and says More) the listener cannot reinforce the mand. In addition, these types of mands are very hard to eliminate once they become strong, and they compete with the development of new mands. Pointing is slightly different and may have advantages for some children. Typical children point frequently as a form of manding. Often this behavior is shaped by providing a choice situation and can be helpful in early manding. Pointing may also be useful for a child who would benefit from a type of picture communication system (see chapter 6). However, pointing has the same disadvantages as More and Please in that a successful mand cannot occur in the absence of the item, and rarely is the point (mand) related to one single item (i.e., a generalized mand). Also, it becomes very difficult to move past pointing and teach a specific word related to each form of motivation. Therefore, when selecting which mands should be taught first, pointing should be avoided as a mand to directly teach. Even if a child comes to the intervention program with a reasonably functional pointing repertoire as a form of manding (but not related to a picture system), soon the focus will need to be shifted to teaching specific words for specific motivators.

Table 5-3

Words for initial mand training that should be avoided if possible

___________________________________________________________________________

Words that are related to a variety of motivators (e.g., More, Please, Mine, Yes, Pointing)

Words that are related to motivators from general categories (e.g., Eat, Drink, Play, Toys)

Words for removing an aversive (e.g., Break, All done, Mine, No, Go play)

Words that are related to items that are hard to deliver (e.g., Swimming, Bike ride)

Words that are related to items that are hard to remove (e.g., Gum, Outside)

Words that are related to politeness (e.g., Please, Thank you)

_____________________________________________________________________________

Words that are related to general categories such as Eat, Drink, Toy, and Play are also problematic in that a number of motivators become related to one response form. For example, Play can be related to jumping, running, self-stimulation, toys, escape from the session, and so on. For some children this type of mand may be appropriate in that one or two mands may be a significant development for a child. However, ultimately it is important that a child acquire specific words for specific items. The problems with mands that are related to a whole class of items, are similar to those described above, in that a single response form may not provide a listener with sufficient information as to exactly what functions as reinforcement for a child at that particular moment. For example, the mand Eat does not let the listener know if a cookie, sandwich, fruit cup, or pop tart is currently motivating. This situation often involves a guessing game on the part of adults, and a delay to the actual delivery of the reinforcer. It is also not uncommon for this situation to evoke negative behaviors. Single class mands may also be hard to change once they have become firmly established in a child’s repertoire.

Another category of mands to avoid in early training are mands related to terminating interaction or tasks, such as All done, Break, Go play or No. These mands may be acquired quickly because they terminate an aversive motivator, but they may directly compete with early attempts to teach a child the important skills he needs. Many children will begin to start saying these words earlier and earlier in a training session, often to the point where they may mand, for example, Break before even presenting the first training trial. On the other hand, for some children who are very difficult learners, it may be possible to establish a first mand as a removal of an aversive stimulus, rather than to obtain a positive stimulus, but this procedure requires careful analysis, intervention, and supervision. Given the choice of developing mands that allow adults to deliver good things to a child, or removing bad things (which often consists of the adult and their related demands), a much stronger relationship can emerge with a child due to the pairing with the delivery of positive reinforcement. When mands to remove aversive (negative reinforcement) are strong in a child’s repertoire, these verbal behaviors often represent major language barriers that must be analyzed and ameliorated as soon as possible (see Chapter 3).

Words that are related to items and activities that are simply too cumbersome to deliver, or are hard to remove, should also be avoided as first mands to teach. For example, a child may be highly motivated by going swimming, but the delivery of swimming as a reinforcer is cumbersome. It can only be delivered once (coming in and out of a pool frequently can become quite aversive), thus is hard to use as a reinforcer to shape a specific behavior (not enough trials). However, as always, an individual analysis of a particular child may reveal that Swimming can be used successfully as a first mand. Certain food items may be hard to consume or remove (e.g., gum). Certain toys may also be hard to remove in order to conduct a sufficient number of training trials, especially if a child has a long history of play with the item in a certain way such as fully assembling a train set. A final category of words to avoid involves those related to social politeness such as Please and Thank you. One simply needs to observe how difficult it is to teach typically developing children to emit these words under the correct circumstances. When they are taught to children with severe language delays they simply become a generalized mand, and have little to do with politeness.

Mand training may be very easy for some children and some of the words identified above may not cause much difficulty. However, more often then not, these are the words that severely limit and obstruct the proper development of a mand repertoire. It is not uncommon to find children with a mand repertoire that solely consists of words from the above list, despite several years of instruction. If these words are avoided from the outset, the probably of developing a more typical mand repertoire is higher. Below are suggestions for what words may work better for early mand training, along with some general guidelines for adding new mand forms to an early learner’s mand repertoire.

Selecting the Best Words for Early Mand Training

Selecting the most desirable first mands for a specific child should begin by making a list of items and activities that are known to function as reinforcement for that child (e.g., cookies, popcorn, juice, bubbles, cars, books, music, Elmo, videos, jumping, spinning). In addition, make note of the circumstance and times when the value of those reinforcers are related to the strongest motivation (Table 5-3). It is important to note that “motivation” and “reinforcement” have different behavioral effects. An item will not function as reinforcement if that item does not motivate a child. Food will not function as reinforcement if a person is not hungry. In its relation to language, motivation occurs before a behavior and evokes a mand (e.g., the desire or motivation for Elmo evokes the mand Elmo), reinforcement occurs after the behavior and strengthens the behavior (e.g., receiving Elmo increases the chances that the child will say Elmo in the future when the motivation is again strong. Although both effects may be related to one item (e.g., Elmo) try and keep them separate by thinking of motivation occurring before a behavior and reinforcement occurring after a behavior. A good way to start looking for early mands is to watch a child and identify items that appear to almost always be valuable or desirable to the child. It is easiest to do mand training on items that have a high motivational value and do not weaken easily. That is why foods and drinks are often good early targets. Be aware of the fact that a child may seem to prefer a desired item if it is free (noncontingent), but if demands are placed on the child the item may quickly lose its value.

Table 5-3

Considering Words to Select as First Mands

___________________________________________________________________________

Words that are related to strong motivation for an item

Words that are related to items that involve slow satiation so multiple trials can be conducted

Words that are related to a specific item (i.e., they involve a nonverbal stimulus)

Words that are related to items that are consumed or dissipate

Words that are related to items that are easy to deliver immediately

Words that are related to items that are easy to remove when necessary

Words that involve a response form is already in the repertoire (e.g., echoic)

Words that do not rhyme other words

___________________________________________________________________________

When selecting the first mands there are several additional variables that can make acquisition quicker. Try and find items that can make use of nonverbal stimuli, that is, there is something visual that can be presented and taken away such as a food item. While tickles and being lifted up may be reinforcing (and may be good first words for some children) it is difficult to make any use of a nonverbal stimulus, thus eliminating a major level of prompting in early mand training. In addition, it may be harder to evoke the mand at specific times, or to include it in a choice array. If specific foods or drinks are used, conduct training after a period of deprivation (i.e., when a child is hungry or thirsty), and deliver very small amounts, such as a single chip, or small piece of a cookie. Small portions will allow for more training trials by delaying satiation (i.e., the point at which the child is no longer hungry or thirsty). If food and drink are both targeted, salty items can increase the value of liquids and, of course there is a well know motivational effect for some people of sugar on milk, highlighted by the “Got milk?” commercials. If toys or other items are reinforcers for a child, try and find items that dissipate or allow for a short period of time such as a wind up toy or music box, bubbles that quickly pop, or a push on a swing. These have advantages in that they also allow for more trials, but they don’t require a trainer to physically remove or take away the fun item (often an aversive event for a child, and unfortunately the trainer is paired with that aversive).

Obviously, there may be only a small number of items that dissipate on their own, and ultimately many of the items that function as reinforcement must be removed at some point in time. For some children removing a reinforcer may not present a problem, but for others it may evoke tantrums and cause other negative behaviors. Certain items and activities can be removed more subtly than others such as using a remote control to stop a video, computer game, or CD. However, many of the items that function as reinforcement for young children (e.g., books, playdough, a slinky, stuffed animals, dolls, crayons, balls, cars, trucks, balloons) must ultimately be removed. While each child differs, a certain time period should be allowed to play with the toy in order to make it fun (5 seconds of a favorite book may be more aversive than reinforcing). Thirty seconds may be better, but the child may lose interest in the item quicker (satiation). Again, it is extremely important to analyze each individual child’s motivation and determine the appropriate contact with a supposed reinforcer to ensure that it indeed does increase the targeted behavior (thus technically functioning as reinforcement).

The items selected for a child should be easy to deliver immediately after the targeted behavior (e.g., a spinning tops, musical books, cars), and last for only a short duration. Items or activities that are delayed (e.g., going to see grandma, going to a movie, going out for ice cream), are difficult to set up (e.g., board games, a block tower to knock down), or take a substantial amount of time to enjoy (e.g., bike riding, swimming, going outside or to a park), should be avoided as early mands. These reinforcers are typically too far removed from the targeted behavior and reduce the number of training opportunities, thus they may not substantially increase the behavior. These forms of reinforcement may still be beneficial for children who are learning other mands, and at some point can still be included in a mand training program, but not as the first mands. However, some children may only be affected by these forms of reinforcement, thus these reinforcers must be used (again pointing out the importance of an individual analysis of each child).

Following the identification of potential reinforcers and an analysis of the motivation related to those reinforcers, another variable to consider is the specific word chosen. Obviously, some words may be harder to pronounce then others, and even though an item may be a strong form of reinforcement, it may be impossible for the child to emit even an approximation to the word early in the intervention (e.g., Elephant). Therefore, it is important to give some careful consideration to the phonetic make up of the targeted word as it relates to the individual child’s vocal abilities. Many nonverbal children can emit some vocalizations, and even perhaps echo a few sounds or words. If this is the case, it will be beneficial to make use of this existing vocal repertoire and attempt to match the targeted mands with the sounds the child can say. If the child has at least some echoic ability, early mands can often be trained in a few minutes. For example, if a child can echo the sound ba and is highly reinforced by watching bubbles float in the air, bubbles would be a good first mand for this child (see the specific procedure below). Even if a child cannot echo any specific sound on command, but makes a variety of sounds, vocal mand training may still be successful. Identify the sound that the child regularly makes and match a few with the child’s reinforcers. For example, if a child regularly is heard to emit the sound Moo and is highly reinforced by watching a movie, this word, motivation, and reinforcer combination would make for a good first mand for Movie. Finally, when considering adding new mands to a child’s existing mand repertoire, avoid words that rhyme with known mands. For example if a child learns the mand Ba ba for bubbles, don’t add new words that start with b like Ball, or Book, even though they may be strong forms of reinforcement, because the child may soon begin to mix them up.

Procedures for Teaching an Early Mand Repertoire

The most important aspect of mand training is knowing what motivates an individual child at the moment of training (current MOs). All mands are controlled by MOs (Table 5-4). If there is no MO for the item or activity at the moment of training, there is no reason to conduct a mand trial. For example, if a child does not want to watch a movie you cannot teach a mand for movie at that point. It doesn’t mean that the child won’t learn other things, but not a mand for movie. He may learn to say Movie, but if there is no MO at that moment it may be more likely it was learned as an echoic, intraverbal, or tact, rather than as a mand. The instructor must determine what functions as an effective form of reinforcement at that moment. In other words, the instructor must be able to identify (tact) the presence and strength of an MO. Often it is quite easy to identify what a child wants, but at other times it can be very difficult.

Identifying a child’s current MO may be difficult for a variety of reasons. MOs vary in strength across time, are often temporary, and may compete with other MOs or the demands placed on a child. For example, a child may first want a specific toy, but when another child has a better toy, the first toy is dropped in preference for the second. Satiation will also affect the value of the MO. For example, popcorn may function as reinforcement for a while, but as the child satiates, the value of the reinforcer decreases. An MO may quickly lose it value if the response requirement placed on the child it too high. For example, a child may have an MO for a cookie, if relatively little behavior is required, but if the required response is high, such as giving up free access to a play environment, a cookie may have little value. MOs may also have an instant or gradual onset or offset. Food deprivation may build up slowly, but decrease quickly. The value of a particular toy may increase quickly and decrease quickly. For example, if a child is given a Magna Doodle® writing board, but no magnetic pen, the value of the pen might be strong immediately. Following a few minutes of drawing, the child might drop the pen and board and move on to another activity. Strong negative behavior may also occur as mands and these must be analyzed and eliminated. Often, the best way to eliminate this type of manding is to teach a more acceptable mand. For example, if a child screams to get outside, the mand Out may quickly eliminate the scream under those circumstances. A language trainer must be aware of all of these varying aspects of the MO, and then know how to bring a behavior under the control of an MO (see below).

Table 5-4

Issues to consider about motivating operations (MOs)

___________________________________________________________________________

All mands are controlled by motivating operations (MOs)

There must be an MO at strength to conduct mand training

MOs vary in strength across time, and the effects may be temporary

MOs must be either captured or contrived to conduct mand training

MOs may have an instant or gradual onset or offset

High response requirement may weaken an MO

Instructors must be able to identify the presence and strength of an MO

Instructors must be able to reduce existing negative behavior controlled by MOs

Instructors must know how to bring verbal behavior under the control of MOs

___________________________________________________________________________

Identifying the Presence and Strength of an MO

How does an instructor know if an MO exists (Table 5-5)? Does the child really want the particular item? Reinforcement surveys may identify what a child liked at some time or another, but they will not tell you if a child wants a particular item at that particular moment. Observation of a child’s behavior in a natural setting can tell you some of what a child wants. Choice procedures that consist of presenting a child with a number of assumed reinforcers (e.g., on a tray) and observing which item is taken might also be effective in determining the relative value of various forms of reinforcement (e.g., DeLeon & Iwata, 1996). They way a child takes an item (e.g., grabs it quickly), interacts with it, and his facial expressions may provided additional measures. If it is an edible, does he eat it quickly, or play with it for a while? Does satiation weaken the mand rate? For example, if a mand is truly evoked by hunger for popcorn, at a certain point most children will become full and no longer want popcorn. If a child keeps on manding popcorn, but does not eat it quickly, or emits negative behavior after it has been delivered, it is possible that the response is not functionally a mand for popcorn (Sundberg, Ramarez, & Garland, 2004). Finally, if the supposed desired item is removed or lost, does a child look for it, or engage in some behavior that demonstrates an MO for the missing item? For a more detailed analysis of the MO the reader is referred to Michael, 2005, and Sundberg, 2004.

Table 5-5

Motivation Operation (MO) Assessment

_____________________________________________________________________________

Test the MO with a choice procedure.

Does the mand occur without verbal or nonverbal sources of control present? (Pure mand).

Is there variation in the mands?

Is the item selected in free access?

Is there a short latency to accessing the manded item?

Is the item consumed or manipulated quickly

Does satiation alter the evocative effect?

Does the mand trial increase a target behavior?

Does negative behavior occur after a mand?

If desired item is removed from choice or menu does the child search for it?

_____________________________________________________________________________

Once an MO has been identified and a response form has been targeted, the goal becomes to teach the child to emit that sound on command. Often the use of a mand frame can help to establish control over a vocal response, even for a child who has never been successful using words to communicate. A mand frame consists of using the three elements of a mand in training: (1) motivation, (2) a response and (3) specific reinforcement. In addition to the mand frame, nonverbal (the object) and echoic (the word) prompts can be used (Table 5-6). The first step is to ensure that the MO is strong (i.e., the child really wants the item). Hold up the item (don’t give it to the child yet), making sure that he wants it (i.e., does he reach for it, look at it, emit other behaviors that have resulted in him obtaining desired items in the past such as whining or grunting). Following this brief MO test, hold up the item again, and say the target word, possibly along with the phrase Say…. For example, if the child has a strong MO for bubbles, the instructor should present the bottle of bubbles (a nonverbal stimulus/prompt) along with the word Bubbles or Say bubbles (an echoic prompt). It is also possible that the verbal prompt What do you want? may be used, but sometimes it may cause problems (see below).

Table 5-6

Using a mand frame and prompts for early mand training

_____________________________________________________________________________

Antecedent Behavior Consequences

Motivating operation (wants bubbles) Praise

Nonverbal stimulus (bottle presented) Ba Blow bubbles

Echoic prompt (Say bubbles)

For some children, prior to requiring a response, it may be beneficial to conduct several noncontingent pairing trials. This procedure consists of (1) using the mand frame and prompts described above, (2) not requiring a response in order to get bubbles, (3) saying the word bubbles prior to blowing the bubbles, and (4) blowing the bubbles. The next step is to begin to impose a short delay between the echoic prompt (Say bubbles) and the pairing of the word (Bubbles!) with the delivery of the bubbles. Depending on the individual child, when it comes time to start requiring a response, it may be that any sound emitted by the child should be reinforced with praise the delivery of bubbles (specific reinforcement). The goal here would be to teach the child that he must say something in order to receive the bubbles. When delivering the bubbles (reinforcement) enthusiastically say Bubbles! prior to blowing the bubbles (the pairing procedure). However, given that bubbles was chosen because of the MO and the fact that the child had said at least ba in the past it would be better to hold out for at least a ba response if possible. If a child fails to show progress (i.e., does not emit any vocal responses), try other motivators and words. If the child still fails to emit vocal behavior, even after a careful analysis and observation by a professional Behavior Analyst, augmentative communication may be appropriate (see Chapters 4 & 6).

For the child who is successful at emitting some vocal behavior, or even better yet, an approximation to bubbles, the task now becomes shaping successive approximations to the target word. If a child is reliably emitting some sound, but not the target echoic response, begin to impose a slightly longer delay to the delivery of the reinforcer. For some children this delay is enough to evoke a better response. The instructor should carefully listen to the topography of the response in order to catch and reinforce closer approximations to the target word. Learning how to shape vocal behavior is often not so easy. Speech pathologists spend many years learning the fine art of listening to speech sounds and identifying which ones should be reinforced and which ones should be ignored (extinction). If a child fails to show progress, consult a professional for advice and help. If a child does begin to emit approximations the target speech sound, the task is still one of shaping closer approximations to the target word. It is not the case that the child must emit the word perfectly before working on fading out the echoic prompt. The issue is function is far more important than form in early language learning. That is, a child should learn to say a correct approximation to several different words before saying a perfectly articulated word for a single item. Typical infants emit “baby talk” for an extended period of time before correctly pronounced words are emitted.

Once a vocal approximation is reliably occurring under the control of the three variables presented in Table 5-6, the next step is begin to fade out the echoic prompt indicated by the dotted arrow in Table 5-7. The goal is to transfer antecedent control from the prompt to object and the MO. This can be accomplished three ways. The first is to begin to impose a delay between the presentation of the object and the delivery of the echoic prompt. For example, hold up the desired bubbles and wait 1-2 seconds before giving the echoic prompt (i.e., Say bubbles.). Interspersing non-delayed trials with delayed trials may help the process of fading out the echoic prompt for some children. Gradually begin to increase the delay to 3 or 4 seconds and provide extra reinforcement for responses that occur prior to the prompt (Touchette & Howard, 1983). The second method for prompt fading is to use partial prompts. The instructor should give only the first part of the echoic prompt (e.g., Say bub…), and differentially reinforce responses that have a form that is better than the partial prompt. The final method is to use a combination of a delay and partial prompt procedure, where the instructor first imposes a delay and then only gives a partial echoic prompt. The goal is to get the child to say bubbles without the instructor saying bubbles. It may be that the verbal prompt Say... plays an important role in letting the child know they must do something. It is possible that the What do you want? verbal prompt has the same function, and could possibly be used at this time.

Table 5-7

Fading out the echoic prompt

_____________________________________________________________________________

Antecedent Behavior Consequences

Motivating operation (wants bubbles) Praise

Nonverbal stimulus (bottle presented) Ba Blow bubbles

Echoic prompt (Say bubbles)

Verbal prompt (What do you want?)

The Correction Procedure and Transfer of Control

When a child makes a correct response and does every thing as planned the consequences are clear--reinforce the behavior. However, when errors occur the intervention is more complicated. The primary goal of a correction procedure is to get a child to emit a correct response with the least amount of prompting necessary, then as soon as possible fade out that prompt. The result should be a transfer of control from one variable (echoic) to another variable (MO/object). In order to accomplish this, the instructor must know at what level of prompt the child is successful (called the operant level), and what is the target level of performance (the immediate goal). For example (Table 5-8), if a child can emit Bubbles when the MO, object, and echoic prompt are all present (Step 1), but not in the absence of the echoic prompt (Step 2), the immediate task is to eliminate the echoic prompt. If the instructor holds up the desired bottle of bubbles and emits Say… (using a delayed prompt procedure described above) and the child fails to respond, the instructor should, after 3-5 seconds re-present the echoic prompt (Step 3). If the child emits a correct response reinforce it with praise, but DO NOT give him the bubbles. Rather, after a very brief delay (2-3 seconds), re-present the original unprompted antecedent (Step 4). This trial is called the “transfer trial.” Many children will immediately repeat the word they just said, if so now deliver praise and the bubbles.

Table 5-8

Correction Procedure for an Early Mand: Phase 1

_____________________________________________________________________

Step 1). Current successful performance (“operant level”)

MO/bubbles bottle/echoic prompt---------------------->Bubbles ------------------>Praise/Bubbles

Step 2). No echoic prompt and an error

MO/bubbles bottle/Say.... ------------------------------>No response ---------------->Praise/Bubbles

Step 3). After a few seconds present the echoic

MO/bubbles bottle/echoic prompt---------------------->Bubbles --------------->Praise/ No bubbles

Step 4). Immediately present the “Transfer trial”

MO/Bubbles bottle/Say…. -------------------------------->Bubbles ------------------->Praise/Bubbles

____________________________________________________________________________

After a sufficient amount of bubble play is provided, and the child still seems to have an MO for bubbles, another trial can be conducted (note that if there is no apparent MO for bubbles, training for bubbles is over, pick another target, or try it later). Present the unprompted trial

diagramed in step 2. If a correct response occurs deliver praise and the bubbles. If an incorrect response or no response occurs repeat the procedure. If the pattern of errors continue, begin to gradually increase the delay between steps 3 and 4. If errors continue, drop back to a delay procedure, plus a partial prompt procedure and work on providing less and less of a partial echoic prompt. If the child can successfully emit the targeted word after a brief delay, then add a distractor trial (Table 5-9). The use of this trial will help to more firmly establish the new source of control. For some children, the early distractor should not involve a vocal response, but rather an imitative, receptive, or matching response. Vocal responses can be tried (e.g., a tact or an echoic), but often on the transfer trial the child will repeat what he just said in the distractor trial. Once the child is successful with one distractor add a second on (e.g., clap hands), then eventually, depending on performance add a vocal distractor. If errors continue to occur, drop back to earlier steps, look carefully at the prompts levels and reinforcement delivery, and assess the relative strength of the MO. Once the child can successfully mand for an item without an echoic prompt efforts should be made to always require that level of responding in his natural environment. However, it may be the case that the response needs “priming” from time to time.

Table 5-9

Correction Procedure for an Early Mand: Phase 2

Adding a distractor trial between steps 3 and 4.

_____________________________________________________________________

Step 3). After a few seconds present the echoic

MO/Bubbles bottle/Echoic prompt------------------>Bubbles ------------------->Praise/ No bubbles

Step 3a). Distractor trial

SD------------------------------------------------>Correct Response------------------------------>Sr

(e.g., touch nose) (touches nose ) (“Right!)

Step 4). Immediately present the “Transfer trial”

MO/Bubbles bottle/Say…. ----------------------->Bubbles ---------------------------->Praise/Bubbles

____________________________________________________________________________

At this point in early mand training, this type of verbal behavior would actually be classified as a mand and a tact (mand-tact), because the response is partly controlled by an MO, but also partly controlled by the presence of the item (this is referred to as “multiple control” by Skinner, 1957). The data are unclear as to whether it is best to add new words or work toward bringing each word solely under the control of the MO (a pure mand). Some research suggests that for some children the quick attempt to remove nonverbal sources of control related to manding may produce rote responses. That is, the responses may not truly be under the control of MOs, rather they may be under the control of other variables such as the verbal stimulus What do you want? (Sundberg, Ramariz, & Garlin, 2004). It also seems to be the case that typical children tend to mand quite extensively for items that are present, thus their responses are at least part mand and part tact. How long this process occurs prior to pure mands in unknown. Therefore, it is recommended that for early learners the instructors keep the nonverbal stimulus present. Early efforts to fade the objects can involve games with occasional trials that consist of hiding the nonverbal stimulus (e.g., put the bottle of bubbles behind the instructor’s back). This procedure may help to begin to free control for the nonverbal stimulus. However, it is important to watch for errors. One common one would be a tendency to call the second targeted item (e.g., popcorn) bubbles when the popcorn is hidden behind the back. In general, it is suggested that prior to removing the nonverbal source of control for the first word, additional mand-tact words be introduced. When adding new words, use the criteria suggested above and the individual child’s performance as general guidelines. For some children, it may be appropriate to introduce new mands while they are still at the echoic level, but for other children if may be important to make sure responding is free from echoic control before introducing new words.

Behavior Modification Versus Behavior Analysis

At this point, it is important to note the distinction between behavior modification and basic behavioral principles (behavior analysis). The procedures described above are only suggestions as to what the behavior modification literature suggests are reasonable strategies to establish a new behavior. The may be several variations of these procedures and there may be other procedures that also may accomplish the same goal. The point here is that there is not just one single way to implementing behavior modification procedures. Clearly there are some ways that work better then others, but it is a mistake to adhere to a rigid formula for prompting, fading and shaping. What is rigid is the basic behavioral principles that underlie these teaching procedures (i.e., reinforcement, extinction, motivation, stimulus control). The goal at this point in the intervention, regardless of how one gets there, is to establish functional control over a vocal response. That functional control should consist of a combination of an MO and a nonverbal discriminative stimulus (stimulus control).

However, a behavioral analysis may suggest that even if the child could now say bubbles under the assumed control of an MO for bubbles and the nonverbal visual stimulus of bubbles, discriminative stimulus and MO control may not exist. In other words, the result of the above procedure could be that the child calls every desired item bubbles. In order to demonstrate stimulus control (a true SD) the must exist a contrasting stimulus condition where the response bubbles is not evoked. This condition is called an S-delta (stimulus-delta). In addition, eventually a second stimulus needs to evoke a different response. In order to establish a discrimination (and thus stimulus control) a second item that is not desired can be occasionally presented (e.g., a shoe). If this stimulus evokes the response Bubbles stimulus control over bubbles has yet to develop. There are a several procedures that may help to establish an SD-S-Delta discrimination. The goal is to teach the child that response Bubbles will only produce the blowing bubbles when the bottle is presented, and not when other items are presented. In order to teach this discrimination, hold up the shoe and if the response Bubbles occurs ignore all responding (an extinction procedure) for a pre-set amount of time (e.g., 30 seconds). Then present the bubbles and if the response Bubbles occurs, blow the bubbles. Then present the shoe again and continue interspersing the two conditions and often differential responding will occur through this procedure.

Another method to establish discriminative stimulus control over a response is to introduce a new stimulus and teach a new response under the control of that new stimulus. For some children this new stimulus can initially consist of interspersing another type of trial such as a motor imitation or echoic trial. This can be easily accomplished by simply presenting a Do this prompt and a known motor response (e.g., clapping) and reinforce the imitative behavior, then hold up the bubbles and prompt with Say…. The goal again is to teach the child that in the presence of one stimulus a certain behavior gets reinforced, and in the presence of another stimulus a different behavior gets reinforced. This procedure also has the added bonus of teaching the child that he must do a little “work” before the opportunity to mands. Note that the focus here is on using the object to develop stimulus control with the assumption that MO control is not yet firmly established. The ultimate goal is differential MO control (i.e., a specific desire evokes a specific word), however, due to the private nature of this source of control (it is often hard to really know what a nonverbal child wants), the recommendation is use the object (the tact part of the early words) as a way to teach the child to emit different words at different times. The task of fading out the nonverbal source of control will be described in a later section.

Adding a New Mand-Tact Word

A third method of establishing stimulus control is to introduce a new mand-tact response. This should be done as soon as possible. The selection of the second word to teach as a mand-tact should be done carefully. The suggestions presented above for the selection of new words applies to the second word with a few additional considerations. The first issue is what MO will be used to establish the response. The MO should be as different as possible from the first targeted MO. For example, if a food item is used first, do not use a second food item (the MO is too similar). Given that bubbles was the example presented above, a food item may be an appropriate new responses. It is also important that the visual stimulus involved be distinctly different from the first stimulus (i.e., don’t use anything that looks like a bottle of bubbles, is the same color, or has the same picture on it). The consequence for the second item should also be significantly different. For bubbles there was a visual moving stimulus as a reinforcer, thus using a spinning wheel as the second reinforcer may be contraindicated. Food, for example, has a consumatory consequence that is clearly different from a visual reinforcer. The selection of the second targeted word is also important. The topography of the response should not be too close to the first word, so in this example it should not start with b or rhyme with Bubbles. The point here is that the probability of establishing a discrimination between two items will be higher if there is a clear distinction between the MOs/stimuli, the responses, and the consequences involved in the first two targeted verbal relations.

Mixed Verbal Behavior Trials (Mixed VB)

A final method of establishing stimulus control and differential responding involves a combination the three procedures described above. This would involve a set of trials that contain the two targeted mand-tact relations, an S-delta condition, and the known imitative, echoic, or matching-to-sample responses. This type of training (“mixed VB”) consists of mixing up the verbal relations (e.g., mand, imitation, echoic and MTS trials). The value of this type of trial, versus those that are mass trials or a specific focus on a certain skill (e.g., imitation or tact only) is that not only is stimulus control strengthened, but eventually it becomes important for a child to be able to respond to a variety of antecedents, some of which may be visual, auditory, verbal, nonverbal or motivate. In addition, a child must be prepared to respond as a speaker or a listener. More advance verbal interactions are typically made up of mixed verbal interactions involving mands, tacts, intraverbals, echoics, etc. At some point in an intervention program a child must be able to respond to this variation of stimulus and motivational variables. The concept of mixed VB suggests that early language learners may benefit from training that involves variation in the presentation of teaching trials.

Data Collection

The type of data that should be collected for the acquisition of the first mands can vary substantially. The primary purpose of collecting data is to identify that learning is occurring, and that the observed learning is a function of the teaching procedures that have been employed. If the data show that learning is not occurring, changes need to be made in the teaching procedures, targeted goals, curriculum, or other variables such as the deprivation levels related to the MOs. The most optimal data would involve measures of the antecedents involved and the response topography on every trial. This level of data collection may only be possible if a second individual is available to record the data. However, in the absence of a second individual, data on the first and last trials of a 15-20 minute teaching session can provide the necessary and sufficient data to provide answers to the issues presented above (Cummings, 2005). First trial data probes consist of recording the child’s response on the very first trial of the teaching session on each specific targeted mand. In addition, it may be helpful to record data on the level of prompt necessary to successfully evoke that mand (a sample data sheet is presented in Appendix 1). For example, on the first trial of the session the instructor should present the child with the bottle of bubbles and the verbal prompt Say…. If the child responds with ba the prompt level of object present should be circled on the data sheet and the response form ba recorded. If the child fails to respond the next level of prompting (echoic) should be presented. If the child responds at this level of prompt the level should be circled on the data sheet, and the topography of the response recorded. Following the first trial probe the correction and transfer procedure described above should be implemented. Following a teaching session of mixed trials, a last trial probe can be conducted. Often, change can be observed in a last trial probe before change is observed in a first trial probe.

Expanding the mand repertoire

Once a child has two mand-tact words (or approximations to those words), a new word should be added. The same criterion suggested above for selecting words should be applied. In addition there should be an increased focus on the other beginning repertoires, specifically, motor imitation, echoic, receptive language and matching–to-sample (see chapters 7-10 for these procedures). The third, mand-tact word should become the target of the intervention and the other mand-tact trials interspersed, along with trials on the other four repertoires. If a child successfully acquires a third mand, and is demonstrating some reasonable echoic behavior, it may be appropriate to begin teaching a few tacts (see Chapter 11). It is important to consider each child individually, rather than applying a rule like “three mands then add a tact.” The goal is to teach several mands and several tacts and each learner will progress at his own rate. Careful consideration and analysis should be given to individual progress and errors when expanding the repertoires.

Generalization

Early on in the intervention program every effort should be made to program for generalization. The point of generalization is that a specific word should occur under a wide variety of circumstances, and not be limited to any single set of circumstances. A simple type of generalization involves conducting the training in different settings. If the child learned the response bubbles in the living room, conduct training in the kitchen, bedroom, outside, at grandma’s house etc. In addition, conduct the training at different times of the day (e.g., morning, afternoon, evening). The noise level and outside distractions should also vary, so the child doesn’t learn to only respond when it is very quite, and no one else is in the room. Then, different people should be able to evoke the mand-tact from the child. If a child has learned to say bubbles as a mand-tact with his mother, his father should require he mand-tact bubbles, as should any significant person in the child’s life (e.g., siblings, teachers, babysitters, friends). The various carrier phrases (e.g., What do you want?” Say.... Look its....) should also be changed if possible, as should the instructor’s tone of voice, pitch, intonation, prosody, and volume of any accompanying verbal stimuli. Finally, and sometimes the hardest form of generalization for children with autism, is generalizing to different materials. Bubbles do not always come in the same size blue bottle with the same label. Change these variables with a variety of different bottles of bubbles. All of the above variables should eventually be combined. For example, after learning bubbles at home with mom before school with a blue 4-inch bottle of bubbles, an instructional assistant at school should ask the child to mand-tact a large red bottle of bubbles in an afternoon group setting, assuming, of course there is a current MO for bubbles. The point is that the response bubbles should be primarily under the control of the target variables, which in this case is an MO and object. The response should not be primarily under the control of an irrelevant variable such as a specific teacher, the color of the bottle, time of day, tone of voice, etc. It is important that generalization be incorporated intro the training program as soon as possible, and with each new word.

Teaching sessions

At this point in the intervention, a teaching session should be short (anywhere from 1 to -15 minutes, depending on the child), and made as fun as possible. Training can be conducted on the floor, at a table, on a bed, playground, or in any conducive environment. The intervention would be essentially the same whether it was conducted in a home or school environment. However, for some children, control of the physical environment may be critical for success. If a child has free access to other reinforcers and adults have very limited instructional control (see Chapter 4), some restrictions (e.g., cornered at a table) may be important. Each brief session may look very similar, with the specific goals relatively straight forward—1) teach the child to mand with as little prompting as possible, 2) teach the child that he must do something (e.g., motor imitation) to get an opportunity to mand, 3) teach the child that the relevant initiating stimulus may be visual, auditory, or both, he may be asked to be a speaker or a listener (mixed VB), 4) gain instructional control over the child, 5) establish the instructor as a conditioned reinforcer (make learning fun, in part by teaching a child that manding gets him desired reinforcers, at a fair price).

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download