289 COMMENTARY

ACHIEVING PARITY: THE ROLE OF AUTOMATIC REINFORCEMENT

DAVID C. PALMER

SMITH COLLEGE

The central insight of Horne and Lowe's article is the importance of the role played by the discriminative effects of one's speech upon one's self. Informed by this insight, Horne and Lowe provide a parsimonious and coherent interpretation of the behavior said to show equivalence relations, exploiting only established concepts of behavior analysis. I am enthusiastic about both the goals and ac­ complishments of the article and therefore will confine myself to suggesting an elabora­ tion or refinement of one or two ideas lightly covered by the authors, particularly the role of automatic reinforcement in the shaping of the speaker's behavior (cf. Skinner, 1957, p. 164, 1979, p. 283; Sundberg, 1980;Vaughan & Michael, 1982).

Speech is special in that we stimulate our­ selves in just the same way and at the same moment we stimulate others. This is a char­ acteristic not entirely shared by sign lan­ guage, because the appearance of a sign var­ ies with the location of the viewer; typically the speaker and the listener view signs from opposite sides. Although subtle, this differ­ ence should impair the acquisition of naming in sign language relative to that of speech, because listener behavior is under the control of stimuli that the speaker can never quite reproduce. Moreover, we might expect more idiosyncrasies, or accents, among signers than among speakers.

The auditory feedback from one's speech can play a special role for a speaker who is already a competent listener. Horne and Lowe allude to the discriminative control that such feedback exerts over orienting behavior and note that this control contributes greatly to our understanding of performance in re­ search on equivalence classes. They also refer to the potential reinforcing function that

Correspondence concerning this commentary should be addressed to the author at the Department of Psy­ chology, Smith College, Northampton, Massachusetts 01063 (E-mail: [email protected]).

such feedback can provide, assuming that the verbal stimuli already function as conditioned reinforcers. A stimulus may be automatically reinforcing if it has been paired with uncon­ ditioned reinforcers; thus, as Horne and Lowe observe, "the sounds and words uttered by parents may function as potent classically conditioned stimuli that have strong emo­ tional effects on the child so that when she hears her own replication of these vocal pat­ terns she generates stimuli that have similarly strong reinforcing consequences" (p. 198). Presumably, then, there may be a reinforcing effect of hearing one's self say, "Good job," because such expressions from parents are likely to serve a reinforcing function.

Horne and Lowe do not make much of this point, and rightly so: We would expect only a relatively few verbal stimuli to function as conditioned reinforcers. Moreover, presum­ ably one quickly discriminates between the praise of another and one's own; the former is far more precious. The reinforcing effec­ tiveness of verbal stimuli quickly becomes conditional on other variables.

However, the feedback from one's own speech plays a different sort of reinforcing function that I will argue is far more impor­ tant in the shaping and development of ver­ bal behavior. One's own utterances can shape and maintain one's behavior, not because of the specific stimulus properties of the verbal stimuli, but because of the parity of such stim­ uli with practices of the verbal community. That is, a speaker who is already an accom­ plished listener can detect when he or she conforms or deviates from typical verbal prac­ tices. Under most circumstances, people find parity of their speech with that of others to be reinforcing and deviations from parity to be punishing. The child who says "Tarry me!" does not have to receive explicit differ­ ential consequences from the verbal com­ munity when she eventually learns to utter "Carry me!" She knows instantly that she has

290 COMMENTARY

achieved parity, because, as Horne and Lowe observe, her repertoire as listener typically precedes her repertoire as speaker.

Achieving parity is a conceptually awkward sort of reinforcer. It is not a stimulus. It is a particular kind of response, a recognition that one has conformed. It is difficult to mea­ sure, even to operationalize; thus, in our in­ terpretations of the acquisition of verbal be­ havior it is seldom given the emphasis it deserves. But, although difficult to measure, the reinforcement is real enough. The effects of achieving, or failing to achieve, parity are particularly conspicuous in endeavors in which social reinforcement is clearly absent or irrelevant. The boy who tries to imitate the sound of a locomotive, an airplane, or a vac­ uum cleaner does not need differential feed­ back from his parents or siblings; the child who picks out the tune to "Mary Had a Little Lamb" on a xylophone may succeed without instruction or approval; the girl trying to learn how to wiggle her ears or to wink profits more from the mirror than from a tutor. In each case, the person already knows what the behavior should look like or sound like, and any behavioral variant that approaches parity is strongly reinforced.

Why is the achievement of parity reinforc­ ing for children? First, it must be acknowl­ edged that it isn't invariably reinforcing in all areas of conduct, as legions of mortified par­ ents in restaurants and supermarkets can at­ test. However, one of the surest ways to opti­ mize one's behavior in a novel situation is to do what others do, and children quickly learn to model their behavior after that of their el­ ders. Nonconformity is often punished with staring, silence, or ridicule. The contingen­ cies for achieving parity in verbal behavior are doubtless subtle, but the embarrassment of those who stutter, lisp, or suffer from other speech impediments suggests that they are nonetheless powerful.

The implications of this source of rein­ forcement are profound. A staple criticism of behavioral interpretations of the acquisition of language, by linguists and cognitive psy­ chologists, is that the reinforcing practices of verbal communities do not seem to be ade-

quate to shape the many subtleties of verbal behavior that children learn to respect. The child who begs to be "tarried" may be car­ ried many hundreds of times without protest; it is not the parents who insist that she get it right. Moreover, as Brown and Hanlon (1970) point out, parents tend to reinforce the content of children's utterances, not the syntax or pronunciation. Despite Moerk's subsequent reanalysis of their data, revealing many sources of reinforcement overlooked by Brown and Hanlon (Moerk, 1983, 1990), the critique is trenchant: Although people re­ spect countless verbal conventions, most of us are unaware of many of them and are in no position to tutor others about them. For ex­ ample, donate and gi,ve are roughly synony­ mous words. One might say, "I gave the mon­ ey to the Jimmy Fund," or "I donated the money to the Jimmy Fund." But, although one might say, "I gave the Jimmy Fund the money," one is unlikely to say, "I donated the Jimmy Fund the money." Our language is re­ plete with such anomalies. To argue that the verbal community explicitly shapes respect for such distinctions would be fatuous. The exquisite subtlety of our verbal repertoires is shaped by the contingencies of automatic re­ inforcement of which Skinner spoke; there are countless such contingencies and they are optimally arranged. To the competent listen­ er, a deviation from parity is instantly detect­ ed; one need not wait for the lumbering ma­ chinery of social reinforcement to swing into action. Identifying units of listener behavior that are relevant to cadence, intraverbal frames, and the other dimensions of syntax remains a formidable problem, but one that is within the scope of the kind of interpreta­ tion pioneered by Skinner and extended here by Horne and Lowe. (See Donahoe & Palmer, 1994, pp. 312-319, for a fuller treat­ ment of this theme.)

The foregoing analysis does not weaken Horne and Lowe's thesis, but it suggests that their interpretation of language development resorts more frequently than necessary to social reinforcers. Second, it suggests that a new mem­ ber needs to be added to their family of effects of a speaker's behavior on him- or herself.

JOURNAL OF APPLIED BEHAVIOR ANALYSIS 1973, 6, 209-217 NUMBER 2 (SUMMER 1973)

SOCIAL CONTROL OF FORM DIVERSITY AND THE EMERGENCE OF NEW FORMS IN CHILDREN'S BLOCKBUILDING1

ELIZABETH M. GoETZ AND DoNALD M. BAER2

UNIVERSI1Y OF KANSAS

The blockbuilding behavior of three preschool girls was analyzed in terms of the forms manifest in any completed block construction, and found to contain few different forms in baseline sessions. Social reinforcement, given contingent on the production of any form not previously constructed within the current session (i.e., every first appearance of any form within a session was reinforced but no subsequent appearances of that form within that session were), increased the number of different forms built per sessions. Social reinforcement, given for all second and later appearances within the session, de­ creased the number of different forms built per session. Furthermore, it was found that new forms (forms never seen before in the child's total prior sequence of blockbuilding sessions) emerged at higher rates during periods of reinforcement of different forms (first appearances) than during periods of baseline or reinforcement of same forms (sec­ ond and later appearances).

Virtually every nursery, preschool, and day care center is equipped with a collection of blocks, usually incorporating a variety of shapes and sizes. Use of these blocks is ordinarily a frequent activity of the majority of preschool children, and such play is widely believed to be educational, contributing to the child's concepts of space, form, mathematics (cf. Cuisenaire rods), balance, leverage, and visual esthetics. The possibility of such a contribution does not appear to have been subjected to experimental analysis and verification, probably because the

1A preliminary version of this study has been pub­ lished in A New Direction fM Education: Behavior Analysis, Vol. I, edited by E. A. Ramp and B. L. Hopkins, (Lawrence: University of Kansas Printing Service, 1971). Supported in part by the University of Kansas Center for Research in Early Childhood Education, in association with the National Program for Research in Early Childhood Education, CEMREL, Inc., and the U. S. Office of Education. (Grant numbers OEC 3-7-070706-3118 and OEC 0-70-4152(607}). We thank Ms. Susan Young, Mr. Charles Burns, and Mr. Lawrence Russell for assis­ tance in obtaining the measurements.

2Reprints may be obtained from either author, Department of Human Development, University of Kansas, Lawrence, Kansas 66044.

outcomes are not well specified in objectively measurable behavioral terms. Even without an empirical basis, though, it might well be guessed that if blockbuilding were to yield such a contribution, the block play itself would need to be diverse rather than limited, repetitive, or stereotyped. If the guess is a good one, then diverse block play becomes a behavioral goal in itself. Significantly, it is one that can readily be given behavioral definition and thereby be sub­ jected to experimental programming, such as by external reinforcement contingencies. If the guess is not a good one, the deliberate develop­ ment of diverse blockbuilding may nevertheless be taken as an experimental goal, so as to ask, if it is developed, what else results? The present study was undertaken to demonstrate the possi­ bility of reinforcing an objectively specifiable aspect of children's blockbuilding that would yield results that might readily be labelled diverse or (less readily, no doubt} even creative. To the extent that such behavior is valued in itself, these procedures then contribute to an educational technology capable of producing it; to the extent that the value of such behavior is questioned, these procedures contribute to the

209

210 EUZABETH M. GOETZ and DONALD M. BAER

possibility of experimental examination of any other results of its deliberate development.

METHOD

Setting

The study took place in a university preschool classroom, specifically in the blockbuilding area of the classroom ( characterized by a clear expanse of flat floor bordered by shelves con­ taining the blocks). Each subject was invited by the teacher to play with the blocks in this area, alone with her and an observer, stationed a few feet away. These invitations were made every few days, and were invariably accepted. The blockbuilding session took place during a time of day when the block area was free of all other children (who were engaged in curricular ac­ tivity outdoors). The session continued until the child said she was finished or until all the blocks available had been used and their rearrange­ ment had stopped, and the child agreed when the teacher asked, "Are you finished?". Usually, the child made a single construction per session; occasionally, several separate constructions were made in a session, in which case all were con­ sidered in the data analysis of that session.

Subjects

Three girls, each 4-yr old, had little else in common apart from an absence of well-devel­ oped blockbuilding skills. This deficit had been remarked on informally by the classroom staff during the months of school preceding the study, and was formally examined in a baseline period at the outset of the study. Staff comments emphasized that these girls' blockbuilding efforts were either devoid of construction ( the blocks merely being laid out in like-shaped or like­ sized groups), or primitive (characterized by repetition of the same structure–e.g., the same "casde"-in every successive construction). One subject came from a low-income family; the other two represented highly enriched, intellec­ tually and culturally stimulating family situ­ ations.

Behavior Definitions and Recording

Child behaviors. The basic child behaviors of blockbuilding were defined according to their products, block forms. In general, form referred to various uses of two or more blocks to create a specified shape or function. Specifically, an arbitrary but frequently seen 20 such forms were defined, as listed in Table 1. Most of these forms could be constructed from a wide variety of blocks; only a few are defined in terms of specific pieces ("circles", "S", "X", and "inter­ face"). Note that some of these definitions re­ quire that the block collection contain rec­ tangular solids of various length-width-thickness ratios and arc-shaped and V-shaped pieces, as well as the more familiar cubes.

A form diversity score was defined as the number of these 20 forms appearing at least once in any session's construction(s).

A new forms score was defined as the number of these 20 forms appearing in a given session's construction that had not appeared in any prior construction by that child (in previous sessions of blockbuilding) recorded within the study. The new forms measure was not scored for the child's first block construction(s) in the first session of the study, in that all forms ap­ pearing in that session would have to be con­ sidered new. Instead, the number of forms appearing in the first construction(s) of the first session of the study was taken as the child's baseline of forms; new forms were then con­ sidered forms other than those and were scored from the second session's construction(s) on.

Recording of each construction was done photographically with a Polaroid camera. A series of photographs was taken of each ses­ sion's construction(s), from all sides and angles necessary for complete display of its structure. (The use of a Polaroid camera was considered important, in that each photograph could be examined for adequacy of exposure, focusing, scope, and clarity almost immediately; deficient ones then could be replaced or augmented by additional photographs before the construction

211 FORM DIVERSITY AND NEW FORMS

Table 1

Definition of 20 Block Forms

FENCE: any two or more blocks placed side by side in contiguity; if not contiguous, then any three or more blocks placed at regularly spaced intervals in a straight line.

STORY: any two or more blocks placed one atop another, the upper block(s) resting solely upon the lower.

RAMP: a block leaned against another, or a tri­ angular block placed contiguous to another, to sim­ ulate a ramp.

PILLAR: any story in which the lowest block is at least twice as tall as it is wide.

POST: any story in which the lowest block is at least twice as wide and half as rall as the upper block(s).

TOWER: any story of two or more blocks, each of which is at least twice as tall as it is wide.

ROOF: two or more slat-shaped blocks placed flat and side by side atop at least two supports.

FLOOR: an inverted roof. BALANCE: any story in which the upper block 1s

at least four times as wide as the lower. ELABORATED BALANCE: any balance in which

both ends of the upper block contain additional blocks.

ENCLOSURE: any arrangement of fences which encloses an open area, with or without "gate".

SUBDIVISION: two or more enclosures in con­ tiguity with at least one common fence.

ARCH: any placement of a block atop two lower blocks not in contiguity.

STORIED ARCH: an arch built atop another arch. ADJUNCT: two or more forms connected by a

fence; at least one of the forms must be an en­ closure, subdivision, or roof.

CIRCLE: arrangement of four arc-shaped blocks in contiguity to form a circle.

"S": arrangement of four arc-shaped blocks in con­ tiguity as two half-circles to simulate an s.

"X": arrangement of two V-shaped blocks in con­ tiguity to simulate an X.

INTERFACE: arrangement of any two blocks with curved contours to fit precisely together, such as circle into hole of doughnut-shaped block, or half­ circle into arc-shaped block.

SIMULATION: a construction of blocks which re­ sembles a real-life object and is explicitly labelled by child as such, usually a building, boat, or swimming pool.

was dismantled). These photographs were taken after the child had left the setting; the camera was not in evidence during the blockbuilding, and the child was presumably unaware of the fact of photographic recording.

Two judges independently examined each series of photographs, counting the number of forms appearing at least once in the series representing the session's construction. Their agreement in this counting of form diversity was compared and found to be 100% over all sessions of the study. In that the new forms score was derived directly from these counts ( comparing the identities of the forms found in a given construction to those found in all previous constructions), there was similarly 100% agreement between the judges on that score as well.

The duration of each session was recorded by the observer with a stopwatch. The session was defined as beginning when the first block was set down, and as ending either when the child

said it was or when the child replied affirma­ tively to a teacher question about completion.

Teacher behaviors. An observer silently watched the blockbuilding interaction from a few feet away and recorded teacher behavior and its contingency with the child's production of forms. Specifically, the observer recorded each successive new form built by the child in that day's construction, whether the teacher attended to that form production or not, whether the at­ tention was enthusiastic and approving, and the duration of the session. These records were used primarily as a check on the teacher's efficiency in carrying out the social contingencies required by the experimental design, rather than as a direct measure of the child's behaviors in producing forms. (The photographs were pre­ ferred as likely to yield higher accuracy, in that they could be examined at leisure many times over, whereas the observer was required to judge all blockbuilding behaviors immediately as they occurred). In addition, the teacher herself main-

212 ELIZABETH M. GOETZ and DONALD M. BAER

tained similar records, as the blockbuilding proceeded. Comparison of the teacher's records and the observer's showed agreement on 95% of the events recorded, over the study, and also showed that 95% of the teacher's attention was supplied in the contingencies required by the experimental design, across the study.

Procedures

Baseline: no reinforcement. During the first three to five sessions (varying with the subject), the teacher sat by the child as she built with the blocks, watching closely but quietly, displaying neither criticism nor enthusiasm about any particular use of the blocks. At the end of each baseline session, the teacher expressed her ap­ preciation of the child's total effort and con­ ducted her back to the rest of the classroom group (usually in the play yard outside), and then returned to photograph the construction(s). Baseline sessions were continued until inspection of the child's daily form diversity scores showed a stable enough level to justify experimental programming. (The new forms score was not considered).

Reinforcement of different forms. After the baseline sessions, the teacher began a program of social reinforcement of new forms. In these sessions, the teacher remarked with interest, enthusiasm, and delight every time that the child placed and/or rearranged blocks so as to create a form that had not appeared previously in that session's construction(s). (The form might well have appeared in any previous session's construction(s), but it would be reinforced in this session nevertheless). In other words, the teacher reinforced every first appearance of any form within the current session but no subse­ quent appearances of that form within that session. The content of the teacher's remarks was designed to accomplish descriptive rein­ forcement: that is, they often indicated the dimension to be reinforced, such as, "Oh, that's very nice-that's different!". These procedures were continued for four or five sessions until clear evidence of increasing form diversity was

obtained, at which point the next experimental condition was implemented. (The new forms score was not considered).

Reinforcement of same forms. Experimental control of the increasing form diversity score was attempted by reversal of the direction of the social reinforcement contingency employed. Thus, for the next two to four sessions, the teacher continued to display interest, enthusiasm, and delight, but only at those times when the child placed and/or rearranged a block so as to create a repetition of a form already apparent in that session's construction(s). (This form need not have appeared in any previous session's construction(s); so long as it had appeared earlier in this session's construction(s), it would be reinforced.) Thus, no first usage of a form in a session was reinforced, but every second usage of that form and every usage thereafter within the session was. Again the content of praise was descriptive: it specified the dimensions (same­ ness) being reinforced (e.g., "How nice-an­ other arch!"). These sessions were continued until a clear decrease in form diversity was seen. (As ever, the new forms score was not considered).

Resumption of reinforcement of different forms. To conclude the experimental analysis and leave each child with the desired high level of diversity, reinforcement of different forms was resumed and continued for three to five sessions, until high levels of form diversity were seen. Procedures during this condition were identical to those used in the previous con­ dition of reinforcement of different forms, i.e., only first appearances of any form within the current session were reinforced.

Number of blocks to be used. The first two subjects, Sally and Kathy, were free to use any number of blocks, many or few, for each con­ struction. In fact, each girl tended to use a number of blocks roughly correlated with her form diversity score, across constructions. While the correlation was not high, it did allow the possibility that the increasing form diversity scores associated with reinforcement of different

213 FORM DIVERSITY AND NEW FORMS

n – NO REINFORCEMENT D – REINFORCE ONLY DIFFERENT FOAMS S- REINFORCE ONLY SAME FORMS

20

10

SALLY J

0"'1———————

IL. 0

20

KATHY

MARY

• • •

I :, z-

10

•• o,_____________________

COOSTRUCTIONS Fig. I. The form diversity scores of three children in the course of block-building training. Initial points,

labelled as n's, represent scores produced when reinforcement was programmed only for different (non­ repetitive) forms; and points labelled as S's represent scores produced when reinforcement was programmed only for repetition of the same forms used previously that session .

214 ELIZABETH M. GOETZ and DONALD M. BAER

forms could represent a chance outcome of the increased opportunity to display different forms when using many blocks. Consequently, Mary's procedures were conducted as described above, but with one added element: Mary was told that she must use all of the blocks (53 10

number) each session. Invariably she did.

RESULTS

Form diversity. Each child's form diversity score, for each experimental session, is shown in Figure 1. It is apparent that the social con­ tingencies applied effectively controlled the development of form diversity: each child showed appreciable increases from her baseline (no reinforcement) levels during the two periods of reinforcement of different forms; and during the interpolated condition of reinforcement of same forms, each child's form diversity de­ creased toward her baseline range.

New forms. Each child's new forms score, for all sessions after the first (which was used to define old forms for all later sessions), is shown in Figure 2, cumulatively. That is, each bar represents the number of new forms displayed by the child so far in her sessions, since the first session. (The number of "new" forms seen in her first session was of course identical to the first point in each graph of Figure 1.) The increment each bar shows over the immediately preceding bar is thus the number of new forms that emerged in that session. Figure 2 shows that the emergence of new forms was largely restricted to periods in which different forms were being reinforced. Only three exceptions appear: Sally, in her second baseline session (first bar in Figure 2), showed one new form, and in her third session, another new form; Mary in her second baseline session (first bar in Figure 2), showed four. By contrast, Sally showed nine new forms during periods of rein­ forcement of different forms, and Mary dis­ played 14 during these periods. In Kathy's case, her 16 new forms emerged only during periods of reinforcement of different forms. In no child's

study did new forms appear during the period of reinforcement of same forms. On the average, across the three children, the rate of emergence of new forms during periods of reinforcement of different forms was 1.5 new forms per session; the rate of emergence of new forms during periods of baseline or reinforcement of same forms was 0.33 per session. These figures ex­ clude the first baseline session, as required by the definition of the new forms score.

Session durations. The mean duration of blockbuilding sessions was 16 min in Sally's case, 19 min in Kathy's, and 9 min in Mary's. However, when Sally was receiving reinforce­ ment for different forms, her mean duration was 23 min; in conditions of baseline and reinforce­ ment of same forms, her mean duration was 9 min. The comparable mean durations for Kathy were 22 min and 12 min, respectively; and for Mary, 10 min and 8 min. Thus, in general, per­ formance during reinforcement of different forms required more time than during baseline or the reinforcement of same forms.

Related Tags