Norm change over subject generations as a function of arbitrariness of prescribed norms

Mark K. MacNeil 
Oklahoma State University
Muzafer Sherif
Sociology Department Pennsylvania State University

Experimental groups making judgments of autokinesis were enculturated with one of two arbitrary norms by planted subjects. Participants in sessions changed through successive "generations" in accordance with the example of Jacobs and Campbell. The principal focus was the effect of differing degrees of arbitrariness on norm formation, persistence, and change. Conformity to imposed arbitrary norms was found to be an inverse function of the degrees of arbitrariness. The two prescribed arbitrary norms were set at clearly separated distances from the natural norm and were clearly differentiated from each other. Both of the arbitrary norms in this study had a much broader latitude (7 inches, or 17.78 cm) than that of Jacobs and Campbell (2 inches, or 5.08 cm). Therefore, as would be predicted from latitude theorizing, the enculturation and transmission of arbitrary norms were more effective and more persistent through generations. The more apparent than real differences between the findings of this study and those of Jacobs and Campbell are reconciled through discussion of the nature of norms.

Since they are the product of interaction among individuals, norms are subject to change as a consequence of further interactions. A norm that forms without the presence of arbitrary influences is a "natural norm" for a given group and the existing conditions. A norm formed under "arbitrary" influences may subsequently be experienced as not fitting the prevailing stimulus conditions, and in consequence, interactions among individuals may lead to norm change (Sherif & Sherif, 1969).

For social units in which the membership changes by means ofa gradual replacement process, as in a family, the norms that are implanted in each successive new member through enculturation are basically derived from the past. Their present form is likely to have been determined, to some degree at least, from past conditions, and such norms are passed on to the new member in the form in which they currently exist, that is, as they have emerged under past conditions. The conditions under which a given norm formed may or may not still prevail. To the extent that the norm being implanted in the new member does not represent the norm that might emerge currently without the influence. of past conditions, the norm implanted in the new member will be arbitrary. The degree of arbitrariness, then, is determined by the difference between the conditions under which the norm originally formed and the conditions (other than the norm's past form) that prevail at the present time. As Sherif and Sherif (1969) have stated:

One important variable affecting the transmission of norms and the extent of conformity to them


by new members is the degree of their arbitrariness relative to current conditions facing the group. When a norm is transmitted by older members to a new generation of members, its "arbitrariness" can be defined relative to the conditions in which the new generation functions. The norm developing in those conditions without an enculturation process by older members can be termed least arbitrary or "natural." (p. 215)

A relevant example of persistence and change is observed in the following account of enlisted men in the United States Army who, prior to World War II, followed an old custom when speaking to an officer: They used the article the and the third person singular of the verb (e.g., "Does the lieutenant want his coffee now?"). Initially, the "old soldiers" imposed this custom on the new selective service "civilian soldiers" with some success. By the time the war had been in progress a year, however, only a few "old timers" continued to use this form of address. As an Army norm the custom had disappeared. Lack of a general societal norm of language honorifics, increased possibilities of moving from enlisted to officer status, and the great influx of nonprofessional soldiers during the war were all conditions that made this norm seem highly arbitrary. Consequently, it did not persist.

An example of an arbitrary norm that persists despite the efforts of some to change it and the considerable awkwardness in its usage is the prescribed spelling of English. This normative system was developed mainly through the attempts of Norman-French scribes to transcribe English words (sounds) using French orthographic conventions. Pre-conquest spellings, the work of Anglo-Saxon scribes unburdened by continental spelling rules, are far simpler and more straightforward. The principal use of the dictionary is to check on the spelling rather than the meaning of words. When we spell phonetically, the form of the words is quite different from "normal" usage. It is evident that our present system of spelling involves rules, that is norms, that are highly arbitrary, yet they persist. The spelling of English words would be vastly different if this normative system, introduced through historical accident, could be removed and the spelling rules allowed to develop under current conditions.

The above examples, and all instances of norm change in actual groups, embody complicated and multifaceted motivational components and established reciprocities, both in terms of organizational structure and normative orientations. The motivational components involved are significant to the members of the cultures and groups concerned. The great distance of our present starting point from the eventual goal of moving experimental analysis to these more complicated areas of persistence and change of norms is clear. In the present study the experimental norms were established without the motivational basis of norms formed in actual groups. The subject's motivation was his desire to do well in the presence of other individuals like himself and to resolve the uncertainties he felt in the situation.

In a study by Jacobs and Campbell (1961) plants (i.e., experimenters' collaborators posing as subjects) were employed to establish an extreme cultural norm in a situation involving the judgment of autokinetic movement. The adoption and survival of the arbitrary norm were then studied as the plants, one by one, were taken out of the group and naive members were introduced, in turn, in their place. The naive members then unwittingly became the further transmitters of the norm to still newer entrants.

Jacobs and Campbell employed six conditions. Two of these conditions were controls, with either one naive subject alone or three naive subjects together forming a natural norm, or baseline, for comparison with arbitrary norms. The remaining four conditions were experimental, and they differed primarily in terms of the size of each generation (from two to four) and the initial number of plants (from one to three). In general, it was found that under all experimental conditions, there was a sharp decay toward the natural norm.

Of particular interest is Jacobs and Campbell's (1961, p. 650) experimental condition "X-4-3" in which they had four-member generations with three plants in the initial generation. The plants presented an arbitrary norm of 15-16 inches (38.10-40.64 cm) under conditions in which the average judgment for the natural norm was approximately

(764) 4 inches (10.16 cm). This was the condition in which Jacobs and Campbell had the highest plant-to-naive ratio and the condition in which plants were in the group over the greatest number of generations. Since only one plant was removed at the end of each generation, it was not until the fourth generation, under this condition, that the four-member group was made up entirely of naive subjects; that is, subjects were exposed to the prescribed arbitrary norm over more generations and, consequently, a greater number of interactions than under the other conditions in the Jacobs and Campbell study. Although Jacobs and Campbell (1961) expected that "the strongest conditions might produce an arbitrary culture which would persist without apparent diminution" (p. 654), they reported that the number of plants had little clear-cut effect.

When the relative structure, that is, the degree of ambiguity, in the different aspects of the total judgment situation are examined, three major areas of consideration become apparent: (a) the structure of the physical stimulus, (b) the structure in the social situation, and (c) the structure in the norm.

The first to consider is the degree of structure in the physical stimulus. Asch's (1951) judgment situation, in which subjects compared the lengths of lines against a standard, did possess a relatively high degree of structure; that is, the subjects could make valid judgments and could match the lines correctly. Asch's subjects did make a high proportion of correct judgments when they were not under social pressure. Further evidence to this point is that despite the uniformity of the plants' judgments in the Asch study, only approximately 37% of the judgments revealed the impact of social influence (Asch, 1951). Looking at the whole situation in which norms may form in relation to a physical stimulus, when the structure of the physical stimulus is relatively high, the potential effect of social influence is proportionately low.

The autokinetic judgment situation provides a physical stimulus that is quite low in its degree of structure. In fact, to determine the nature of the stimulus in terms of the range of judgments, a consensual base must be determined as is the case with psycho- social scales, because there is no physical reality. The only "reality" is the subjectively based reality of the individuals' span of judgments, which does establish certain limits. The light moves only subjectively; there is no physical movement. The ease with which social influence shifts subjects away from the natural range of judgments is further evidence of the relative lack of structure of the autokinetic stimulus situation. When the physical stimulus lacks structure, one confident plant would be expected to be able to cause distinctive, if only temporary, shifts in the perception and judgments of naive subjects, and such is the case with autokinesis (Goldberg, 1954; Kidd, 1958; Pace & MacNeil, 1974).

Why, then, did the arbitrary norm not form and persist, as prescribed even under the conditions of greatest social influence in the Jacobs and Campbell (1961) study? A closer examination of the location and structure in the situation is revealing. If the degree of structure of the physical stimulus is low and the structure with regard to social influence is high, and there is still a failure of the social norm to form at a prescribed degree of arbitrariness and to persist, the problem must lie in the structure of the arbitrary norm itself, its reasonableness under the conditions, its credibility.

The nature of a norm involves more than simply its central tendency. Most social norms include a latitude, a range of acceptable percepts or behaviors and are seldom an all-or-nothing punctate (Sherif & Sherif, 1969, pp. 192-196). With regard to autokinetic judgment, there is invariably a range of judgments around a model tendency. When the central tendency of the judgments made in a particular autokinetic situation falls at 4 inches (10.16 cm; approximately what Jacobs and Campbell, 1961, reported for their "natural" condition), the range of the norm is approximately 7 inches (17.78 cm); that is, the latitude of the norm is bounded by judgments of 1 inch and 7 inches (2.54 and 17.78 cm). Only approximately 2.5% of the judgments made fall beyond these bounds.

( 765)

In view of the concept of a norm as a reference scale, the norm that Jacobs and Campbell prescribed, as given by their plants at 15-16 inches (38.10-40.64 cm), is not only highly arbitrary in terms of its locus, but it appears to be extremely unrealistic in terms of its latitude. It is small wonder that upon removal of the plants from the group, this norm neither effectively transferred to the naive subjects nor persisted over generations of membership change. The inherent structure of the prescribed norm, per se, was more than sufficient to overcome the ambiguity inherent in the autokinetic situation.

Further, Jacobs and Campbell's 8-foot (2.4-m) subject-stimulus light distance additionally constricted the lability of apparent movement. The subject-stimulus light distance in the present study was 16 feet (4.8 m). The effects of various stimulus factors in the autokinetic situation are revealed in the studies of Sherif and Harvey (1952) and Hood and Sherif (1962). Among the factors involved are the duration of the stimulus light presentations, the subject-stimulus light distance, the intensity of the stimulus light, the subject's subjective impression of the size of the laboratory area, and so on.

The process of reversion to the natural norm through new generations offered by Jacobs and Campbell (1961) is a plausible one:

The deterioration of the cultural norm occurs, in part, because each naive respondent makes judgments lower, somewhat closer to the natural norm, than do confederates or elder citizens serving as his mentors (p. 655).

This explanation certainly sounds reasonable and is undoubtedly the case when there are factors such as the natural norm in the autokinetic situation or factors that are perceived by members of new generations as reasons for change in real social units. When we observe the persistence of apparently arbitrary norms in our own and other cultures, we are reminded that the real-life condition resembling the "natural norm" of autokinesis frequently does not exert sufficient influence to change the norm. Norms often remain in arbitrary states over many generations. The

stimuli to which social norms pertain are, of course, often highly unstructured. The referents of some social norms must be highly unstructured for such norms to persist at the highly arbitrary levels at which we find them. The experimental situation in which judgments of autokinesis provide social norm prototypes, as used by Sherif (1935) and others, can offer a set of conditions analogous to real life (Hood & Sherif, 1962; Sherif, 1935; Sherif & Harvey, 1952; Walter, 1955; Whittaker, 1958).

An extension of the present line of research dealing with the fate of arbitrary norms, or traditions, is relevant. Varying two aspects of arbitrariness in an experimental situation involving a game, Weick and Gilfillan (1971), by manipulating the rules of procedure and the difficulty of playing according to those rules, extended the Jacobs and Campbell paradigm to a less structured and relatively more natural interpersonal interaction format. Male subjects were placed in social units of three members each. They were prohibited from communicating with each other in the course of determining what number each subject would present in order that their individually presented numbers (limited to digits between 0 and 10) would add up to match a target number, which was given by the experimenter. The procedure used for accomplishing this, the goal of the game, constituted a procedural norm for the social unit. Any single procedural norm was considered

"arbitrary," since there were at least 25 procedures that could be used. However, some procedures were easier to follow than others. The difficult procedures involved rules that were more complicated and required a greater degree of formalization than the easy procedures.

Weick and Gilfillan (1971) assigned either easy or difficult procedures to the three subjects making up the initial generation of each experimental social unit. Each new member added to the social unit was given instructions by the experimenters as to the specific procedures for the unit, and an "old" subject was removed as the new member joined the unit. This procedure was followed to generate 11 generations of subjects playing the game

( 766) according to the normative procedures for their own social unit.

In the Weick and Gilfillan (1971) study, it was found that arbitrary procedures that were difficult and complicated were changed to easier procedures by the fourth generation. Arbitrarily assigned procedures that were easy were adhered to throughout the experiment, although a number of equally easy procedures were available.


On the basis of cultural evidence and experimental findings, hypotheses were advanced concerning norm persistence and change in judgments of the extent of autokinetic movement.

In terms of both their latitude and focus, norms will persist for shorter or longer durations, depending on the degree of arbitrariness under which they are established. Specifically, the following five predictions were made:

1. A norm formed in the autokinetic situation without experimental enculturation (a natural norm) will be transmitted to succeeding generations and will persist with minor variations.

2. When a majority of individuals in an experimental group give judgments within a prescribed range and around a common mode (experimental enculturation), new participants will adopt their norm, given that the enculturation norm is not extremely arbitrary for the defined conditions of norm formation.

3. When the cultural agents (planted subjects) are removed one at a time in successive generations, the extent of transmission of the enculturated norm to the just-succeeding generation will vary inversely with the degree of arbitrariness of the enculturated norm.

4. As individual participants change over successive generations, there will be a general shift from the more arbitrary enculturated norms toward the least arbitrary (natural) range and mode.

5. The persistence of the enculturated norm over succeeding generations with changing participants will vary inversely with the degree of its arbitrariness; conversely-. the change from the enculturated norm will vary positively with the degree of arbitrariness of the enculturated norm.


Testing these hypotheses required the establishment of norms under different degrees of arbitrariness and the study of their persistence and change through successive subject generations, holding other conditions constant.

The study was designed to test the hypotheses in the autokinetic situation, varying the degree of arbitrariness of the norm to which the individuals were exposed (enculturation norm). The implantation, or enculturation, process occurred during the interaction of four subjects who gave vocal estimates of the distance of apparent movement. In a given session, following each series of 30 judgments, that is, each generation, the participants were successively changed by the removal of one experienced subject and the introduction of one naive subject (Jacobs & Campbell, 1961).

The autokinetic equipment employed is described in Sherif and Harvey (1952; pp. 284-285) and Hood and Sherif (1962). The experimental room was 14 X 35 feet (4.2 X 10.5 m) with subject-light distance set at 16 feet (4.8 m). Exposure time of the light was 3 sec, and provision was made for four subjects. In the course of each generation, 30 judgments were voiced aloud by each of the four subjects. At the end of each generation, there was a 5-minute rest period under extremely dim red illumination. During the rest period, the old subject, or plant, was removed and the new subject was introduced.

The least arbitrary, or natural, norm (Condition 1) formed during the interaction of naive subjects throughout successive generations. Thus, the enculturation norm was produced by the first generation of naive subjects. In the first generations, under Conditions 2 and 3, there were three "planted" subjects, who gave arbitrary estimates within a range and around a mode prescribed by the experimenter, and one naive subject. The plants were removed successively during the first three generations, a naive subject being added each time. All subjects were naive by the fourth generation.

In Condition 2 (moderately arbitrary), the planted subjects gave estimates within the range of 9-15 inches (22.86-38.10 cm) around a mode of 12 inches (30.48 cm). In Condition 3 (most arbitrary) the prescribed norm range was 15-21 inches (38.1053.34 cm), and the mode was 18 inches (45.72 cm). In both of these conditions, in the course of each generation, each planted and naive subject voiced 30 judgments of the extent of movement. The prescribed judgments given by the planted subjects were given in an apparently random order; the actual prescribed distributions resembled a normal curve with the focus of judgments at 12 and 18 inches (30.48 and 45.72 cm), respectively.

In Condition 1 (natural norm), the process of formation and transmission of the natural norm was

( 767) studied through eight generations of subjects. In Conditions 2 and 3, the enculturation process lasted for three generations, and the transmission by naive subjects continued for another eight generations. Two replications of each condition were conducted.

Subjects and "plants" were male, high school students between 16 and 19 years of age. Control requirements were that no subject know, or have an established social relation with, any other subject in the experimental situation. Subjects were told that the purpose of the experiment was to find out how accurately people could estimate the distance a light moved in the dark. The need for this information by military and space programs was stated.


The concept of norm, defined as a latitude of behavior within which exists a focus of behavior, requires that a measure of conformity and a measure of typicality be used to fully represent the persistence and change of norms as they are transmitted across succeeding generations. Analysis of the changes of the norms that occurred, predictably, as a function of the degree of arbitrariness of the original norm, was made using both conformity and central locus measures. Conformity to the prescribed latitude under each condition, that is, the degree of arbitrariness, was defined

as the proportion of judgments made by the naive (new) member of each generation that fell within the originally established normative range. For the moderately and most arbitrary norms (Conditions 2 and 3, respectively), latitudes were established during enculturation, with the plants giving judgments, from 9-15 inches (22.86-38.10 cm; Condition 2) and 15-21 inches (22.86-53.34 cm; Condition 3).

Generation Range: (25th - 75th percentile) Mdn M
1 4-8 (10.16 - 20.32) 6.0 (15.24) 6.0 (15.24)
2 2-7 (  5.08 - 17.78) 5.5 (13.97) 5.2 (13.20)
3 3-5 ( 7.62 - 12.70) 4.0 (10.16) 3.8 (9.65)
4 2-6 (5.08 - 15.24) 4.0 (10.16) 4.1 (10.41)
5 3-6 (7.62-15.24) 4.5 (11.43) 4.2 (10.67)
6 3-6 (7.62-15.24) 4.5 (11.43) 5.0 (12.70)
7 3-7 (7.63 - 17.78) 4.5 (11.43) 5.1 (12.95)
8 3-6 (7.62 - 15.24) 4.5 (11.43) 4.6 (11.68)
Note. In each generation there were four naive subjects making 30 judgments, with two replications
a All judgments were given to the nearest inch. Numbers in parentheses give metric conversion

The natural, or least arbitrary, norm (Condition 1), which emerged with a range of 2-8 inches (5.08-20.32 cm), matched in latitude the range of the moderate and most arbitrary conditions, that is, a width of 7 inches (17.78 cm; see Table 1). Since subjects were randomly assigned to both the conditions of arbitrariness, and to the generations within each condition, a completely randomized factorial analysis of variance was applied to the proportions following transformation to arc sine radians (Kirk, 1968; see Table 2).

Norm persistence and change, as reflected by the central tendencies of the norms that emerged under the differing conditions of arbitrariness, were assessed by examining the medians of the judgments of all four of the subjects in each generation. Since a social norm in its entirety consists of the judgmental range, which of necessity includes the foci, of all members of a social unit, it is this norm in its entirety to which the new member

    Proportionsa Medians
Source df MS F MS F
Levels of arbitrariness (A) 2 17.435 53.577*** 115.842 64.409***
Transmission generations (B) 7 0.873 2.683* 19.332 10.916***
A X B 14 4.38 1.345 5.918 3.342**
     Within-cell error 24 .325   1.771  
a Transformed to arc sine radians
* p < .05
** p < .005
*** p < 0001.

(768) is exposed. The courses of the norms across generations were then studied in terms of the central tendencies of the norms by a completely randomized factorial analysis of variance, using the medians of judgments as the basic data representing the foci of judgments for each generation. The factorial analysis was the most conservative assessment of effects applicable, since the design required taking into account the complex pattern of the interrelationships of the subjects.

Hypothesis 1. As predicted in the first hypothesis, the norm formed under the least arbitrary, or natural, condition (Condition 1), with no planted subjects in the situation, was transmitted to succeeding generations of subjects with minor variations. In terms of conformity to the latitude of the norm, the proportions of judgments falling within the original range of the natural norm (in arc sine radians) showed the least variability across generations of the three conditions, that is, levels of arbitrariness as evidenced by the simple main effect test, F(7, 24) = .523, ns (see Table 2).

In terms of the focus of judgments, represented by the average of the medians of judgments of all four subjects for each generation (see Table 3), Condition 1 was the least variable of the three conditions of arbitrariness across generations, with a simple main effect test, F(7, 24) = .848, ns (see Table 2).

In terms of both the transmission of the focus of the norm and the adoption of the range of the norm by the new subjects in each generation, it is evident that the least variability of the three levels of arbitrariness across transmission generations was found in Condition 1, the naturally established norm.

The natural norm in this study was formed without the imposition of a prescribed arbitrary norm and was transmitted through successive generations. It provides the baseline to evaluate the persistence and change of norms produced under conditions consisting of the imposition of arbitrary norms.

Hypothesis 2. Figure 1 and Table 3 present the percentage of judgments within the prescribed ranges for Conditions 2 and 3, through three enculturation and eight transmission generations. The second hypothesis, that under the more arbitrary Conditions 2 and 3, the naive participants in the enculturation generations would adopt the range and mode used by the planted subjects, was supported, the extent of this conformity being relatively greater in Condition 2 than in the most arbitrary Condition 3. In Condition 3 (most arbitrary), over 90/0 of the naive subjects' judgments were within the prescribed range during the first two enculturation generations and 79 % in the last when only one planted subject remained. In Condition 2 there was almost complete conformity to the moderately

Figure 1 Conformity to enculturation norms
Figure 1 Conformity to enculturation norms in the autokinetic situation. (Prescribed norm for Condition 2, moderately arbitrary, is 9-15 inches or 22.86 - 38.10 cm. Prescribed norm for Condition 3, most arbitrary, is 15-21 inches, or 38.10-53.34 cm.)


Table 3 Means, Medians and Percentage of Judgments For All Subjects and For Each Naive Subject With Prescribed Range by Generation for Conditions
Condition Condition 2a   Condition 3b
M Mdn % within all Ss % within naive S   M Mdn % within all Ss % within naive S
E1 11.8 11.7 98 99   17.4 17.0 95 95
E2 11.3 12.0 94 90   17.6 17.5 92 91
E3 11.8 12.0 97 95   16.9 16.5 79 69
T1 11.2 11.2 91 93**   15.9 16.5 63 55**
T2 10.6 10.5 83 82***   13.1 12.5 28 23***
T3 10.5 10.2 74 60**   10.6 10.5 18 15**
T4 9.9 9.5 72 70****   8.2 7.7 5 5****
T5 9.9 10.0 65 40**   7.6 7.5 2 1**
T6 8.2 8.2 37 29*   7.0 6.5 1 0*
T7 8.1 8.1 37 38**   7.1 7.0 1 0**
T8 7.3 7.5 27 25*   6.5 6.0 2 0*
Note. For both conditions, n=2 per cell.
a 9-15 inches, mode = 12 inches
b 15-21 inches, mode = 18 inches
* p < .10
    ** p < .05
  *** p < .01
**** p < .005

arbitrary prescribed range throughout the enculturation generations.

Figure 2 and Table 3 present the medians of naive subject judgments, by generation, for enculturation and transmission generations. This measure of norm typicality confirms the evidence of conformity to prescribed norm latitude by indicating that subjects in Condition 2 distributed their judgments around medians that were very close to the prescribed mode of 12 inches (30.48 cm) during the enculturation phase. Naive subjects in the enculturation phase of Condition 3 (most arbitrary) centered their judgments within an inch of the prescribed mode of 18 inches (45.72 cm) in the first two generations, and within an inch and a half in the third generation. Such variation is well within that found for the least arbitrary, or natural, condition.

Hypothesis 3. Following enculturation, after all planted subjects had been removed, the extent to which the prescribed norm was transmitted to the just-succeeding generation varied, as predicted, with the degree of arbitrariness of the enculturation norm. In Condition 2, approximately 93.5% of the judgments of the naive subjects in the first transmission generation fell within the prescribed range as compared with approximately 55% in the most arbitrary Condition 3. This difference is statistically significant, t(24) = 1.8748, p < .05. The results show that enculturation had taken place and that the initial transmission occurred to an extent depending on the arbitrariness of the prescribed norm. Thus, to this effect, Hypothesis 3 is supported.

Hypothesis 4. Hypothesis 4 predicted a general shift from the norms initially established (enculturated) in Conditions 2 and 3


Figure 2
Figure 2 Averages of judgment medians of replications. (There is no prescribed norm for Condition 1, natural. The prescribed norm for Condition 2, moderately arbitrary, is 9-15 inches or 22.86 - 38.10 cm, with a mode of 12 inches, or 30.48 cm. The prescribed norm for Condition 3, most arbitrary, is 15-21 inches, or 38.10-53.34 cm, with a mode of 18 inches, or 45.72 cm.)

toward the least arbitrary, or natural, range and mode over successive generations. The shift of percentages within the prescribed ranges, as shown in Figure 1 and Table 3, as well as the medians shown in Figure 2 and Tables 2 and 3, strongly support this hypothesis. Differences among the medians by simple main effect tests under all three conditions of arbitrariness are significant up until Transmission Generation 8, F(2, 24) = 2.541, p > .10, ns.

Hypothesis 5. The distinctive prediction of the present research was that persistence of an enculturated norm over succeeding generations, with changing participants, would vary inversely with the degree of arbitrariness of the prescribed norm. Conversely, the trend of change toward the natural norm would vary directly with the arbitrariness of the enculturated norm. This prediction was supported.

A comparison of the shifts of the means of the judgments made under Conditions 2 and 3 (moderate and most arbitrary) by Transmission Generations 1-8 indicates a more rapid change toward the natural norm under the most arbitrary Condition 3 than under the moderately arbitrary Condition 2 (see Figure 2 and Table 3). Until Transmission Generation 4, when under the most arbitrary condition the focus of the norm leveled out, the rate of change under the most arbitrary condition is conspicuously greater. By Transmission Generation 3, 56.4%of the total change in the mean that occurred in Condition 3 (most arbitrary) had already taken place. By Transmission Generation 4 in Condition 3, 81.9% of the downward change in the mean had occurred. In the moderately arbitrary Condition 2, only a third of the total change occurring in that condition had been accomplished by the fourth generation. It was not until the sixth generation in Condition 2 that as much as 50,%o of the total shift occurred. By the sixth generation, 77% of the change occurring in the means throughout Condition 2 was completed. From Table 3 and Figure 2, it can be seen that the median of judgments was higher in Condition 2 from the fourth generation onward than in the most arbitrary condition (Condition 3).

( 771)

In terms of conformity, the effect of the degree of arbitrariness on the persistence and change of the norms is equally apparent. Figure 1 and Table 3 indicate that by Transmission Generation 4, the norm under the most arbitrary condition (Condition 3) has only 5%0 of the judgments within the prescribed norm and continues to contain less than 3% of such judgments through Transmission Generations 5-8. Under the moderately arbitrary condition (Condition 2), the norm still has over 710 of the judgments within the prescribed range during the fourth generation, and even in the eighth generation it has above 26% within the original arbitrary range. The greater persistence of the prescribed norm in Condition 2 is also apparent in Table 3 and Figure 1, which show the extent of conformity to the prescribed ranges.

Newman-Keuls tests of the differences between means of the transformed proportions for the different conditions of arbitrariness, across generations, were computed. The mean of the transformed proportions for Condition 3 (most arbitrary) was .524, for Condition 2 (moderately arbitrary) it was 1.752, and for Condition 1 (least arbitrary-natural) it was .60. Comparing Condition 1 with Condition 3, the observed q(3, 24) = 5.147, p < .01. Comparing Condition 2 with Condition 3, the observed q(2, 24) = 3.044, p < .05. The implication of the order of the means of the different levels of arbitrariness supports the prediction that while enculturation was effective, the total amount of judgments within the range, hence the differential rate of change, was ordered from the greatest persistence in Condition 1 to the least persistence (greatest change) in Condition 3.


An understanding of the procedures followed by Jacobs and Campbell (1961) and the critical modifications in the nature of the prescribed norms made for this study is essential for understanding the differences in the outcome of each. Jacobs and Campbell presented an ingeniously elegant experimental design, which although lacking the highly confounding inputs found in norm formation during the normal socializing process, does provide a straightforward and accurate laboratory analogy to the formation and change of norms over the membership generations found in social units such as families and groups.

Jacobs and Campbell used planted subjects to inculcate a norm. The latitude of their prescribed norm, 15-16 inches (38.10-40.64 cm) was, however, excessively arbitrary under conditions in which a mode of approximately 4 inches (10.16 cm) was natural. When this arbitrariness of latitude was combined

with the arbitrariness of a prescribed mode at approximately 15 inches (about 38 cm), an unrealistic level of extreme arbitrariness resulted. In terms of published data on norm formation in the autokinetic situation, and particularly in the small laboratory room (8-foot, or 2.4-m, subject-stimulus distance) used by Jacobs and Campbell, this prescribed range and mode was so extremely arbitrary for the conditions, that the subjects neither accepted fully nor retained the norm. Analogies in the cultural area are difficult, since almost all cultural norms permit some individual variations, whereas 1 inch in the range of autokinetic judgments amounts almost to

no variation psychologically. The results, not surprisingly, were a lack of compliance to the prescribed norm, a lack of persistence of the somewhat lower norm that did form, and a rapid collapse of that norm.

The results, as reported by Jacobs and Campbell (1961), suggest that their initial enculturation attempt was not highly effective. Therefore, the predictions in the present study concerned the formation of the norm during the enculturation process as well as the effect of arbitrariness on the norm as it persisted and changed.

It was predicted and found, in both the Jacobs and Campbell study and our experiment, that the norm formed under the least arbitrary, that is, the natural, condition was transmitted to succeeding generations of subjects with minor variations and with an increasingly greater degree of stabilization. Under the conditions of greater arbitrariness (our Conditions 2 and 3), the naive participants in the enculturation generations did

( 772) adopt the range and mode presented by the planted subjects, the extent of their conformity being relatively greater in the moderately arbitrary condition (Condition 2) than in the most arbitrary condition (Condition 3).

Following enculturation, when all planted subjects were removed, the extent to which the prescribed norm was transmitted to the just-succeeding generation of naive subjects varied inversely with the degree of the arbitrariness of the enculturation norm. Thus, given the fact that Jacobs and Campbell's condition represented a greater degree of arbitrariness (because of its very constricted latitude) than our most arbitrary condition (Condition 3), the results in both experiments are predictable by the following proposition: The degree of enculturation of an experimentally introduced norm (of low degree of self-involvement), and its transmission through succeeding generations, is an inverse function of the degree of its arbitrariness.

It is noteworthy that using an altogether different, but conceptually equivalent setup, Weick and Gilfillan (1971) confirmed the validity of the proposition. The task of the subjects in their experiment was more concrete and, perhaps, more challenging than just estimating the distance a point of light was perceived to be moving in complete darkness.

In social life, in the actuality of normative and functional processes of human groups, the norms are not merely transitory events in time. Their rise, change, persistence or decay over time, and transmission through generations, are understandable only when the conditions of their formation and the course of their natural history, under various circumstances, are included in the account. Therefore, for the following reasons it can be deservedly acclaimed that Newcomb's Bennington research (Newcomb, Koenig, Flacks, & Warwick, 1967) will survive the test of time as a significant landmark in social psychology: (a) Newcomb specified the conditions of interaction (in the newly founded, New Deal permeated Bennington college community) that brought about changes in the norms and values of the participants; (b) he followed the fate of these changes-or the resistance to them-in the participants, over time, through the senior year of the college generation; (c) going far beyond that, Newcomb and his associates assessed the persistence---or change---of these norms over a 20-year period of time.

If the social psychologist is to broaden his perspective (as he should) in the study of the internalization of norms and their fate over years and generations under various circumstances, he must either turn to the historian's works or become a historian himself. Under the constraining conditions of the laboratory, the internalization of norms and their persistence and change can be' studied only as microsimulations of the actualities of the problem area. In a real sense, every laboratory experiment is (and should be) a simulaion. But simulation is a most difficult undertaking; microsimulation of situations has to embody in miniature the essential properties of the actualities of the problem studied. A simulation study is not simply a "fancy" technique, although uncritically it may be treated as such; it has degenerated to that level, in various quarters in recent decades. Lack of concern with the issue of the isomorphism between the technique used in simulation and the actualities of a problem area is nothing less than scientific irresponsibility, no matter how impressive the results of such nonchalance may appear.

The only justification for our own and similar laboratory situations in the study of norms is their essential, common structural characteristics, which are as follows:

1. A fluid, unstructured stimulus situation exists that allows for various alternatives (within limits) for the distribution of judgments.

2. This unstructured stimulus situation results in the subjects' distributing their judgments without experimental prescription (Condition 1) into a normative latitude of their own within the limitations set by conditions of the setup (distance, intensity, duration of the light, etc.). During this process, they (consciously or unconsciously) influence each other with respect to the task at hand. The erratic distribution of responses is discomfiting to the participants, while stabiliza-

( 773) -tion is settling for them. Thus, in this highly ambiguous situation, where there are no sure external cues (reference points) to facilitate confident judgments, participants rely more on each others' judgments. This dependency on social cues increases proportionately to the degree of unstructuredness (uncertainty) in the situation.

3. With the experimental prescription, the subjects can be psychologically pulled to the prescribed latitude during the actual application of influence (enculturation sessions), and the influence persists for some generations if the prescription (as in Condition 2) is not too much removed from what they by themselves would settle for under this circumstance. However, if the external prescription is too far off (as in Condition 3), subjects do not completely accept the norm as it is presented, even during the application of influence, and the effect of this influence is transitory, as evidenced by the more rapid collapse of the prescribed arbitrary norm in the course of the transmission sessions.

It cannot be argued that the task in our experiment even approximates the self-involvement of the subjects, as might the issues of politics or social reforms, which may be studied in field research of attitude change and persistence. In self-involving issues (like politics or religion), attitudes tend to persist for long periods and may persist through generations, even when they become maladaptive relative to changing objective conditions. This was illustrated earlier in the article. More striking illustrations can be readily selected from political, religious, and social areas. An area of research concentrating on the problem of persistence of attitudes and myths that have become dysfunctional-maladaptive in terms of prevailing conditions at present-may well be one of the needed developments in social psychology. We consider the line of research represented by Jacobs and Campbell (1961), Weick and Gilfillan (1971) , and our studies as a baseline start along this area. The fruitfulness of this line of research can be established only by experimentally demonstrating that it can be extended to handle situations (tasks) that are considerably self-involving for the subjects. If this difficult feat can be accomplished, we will be gaining the needed tie-in between the unobtrusive type of laboratory rigor and the close-to-life realism of field studies (such as represented by observation, interview, and attitude validation through assessment of actual behavior).


Asch, S. E. Effects of group pressures upon the modification and distortion of judgments. In H. Guetzkow (Ed.), Groups, leadership and men. Pittsburgh, Pa.: Carnegie Press, 1951.

Goldberg, S. C. Three situational determinants of conformity to social norms. Journal of Abnormal and Social Psychology, 1954, 49, 325-329.

Hood, W. R., & Sherif, M. Verbal report and judgment of an unstructured stimulus. Journal of Psychology, 1962, 54, 121-130.

Jacobs, R. C., & Campbell, D. T. The perpetuation of an arbitrary tradition through several generations of a laboratory microculture. Journal of Abnormal and Social Psychology, 1961, 62, 649-658.

Kidd, J. S. Social influence phenomena in a task-oriented group. Journal of Abnormal and Social Psychology, 1958, 56, 13-17.

Kirk, R. E. Experimental design: Procedures for the behavioral sciences. Belmont, Calif.: Brooks/ Cole, 1968.

Newcomb, T. M., Koenig, L. E., Flacks, R., & Warwick, D. P. Persistence and Change: Bennington College and its students after twenty-five years. New York: Wiley, 1967.

Pace, D. J., & MacNeil, M. K. Comparison of social norms across four judgment situations. Perceptual and Motor Skills, 1974, 39, 575-582.

Sherif, M. A study of some social factors in perception. Archives of Psychology, 1935, 27, No. 187.

Sherif, M., & Harvey, 0. J. A study in ego functioning: Elimination of stable anchorages in individual and group situations. Sociometry, 1952, 15, 272-305.

Sherif, M., & Sherif, C. W. Social psychology (Rev. ed.). New York: Harper & Row, 1969.

Walter, N. A study of effects of conflicting suggestions upon judgment of the autokinetic situation. Sociometry, 1955, 18, 138-146.

Weick, K. E., & Gilfillan, D. P. Fate of arbitrary traditions in a laboratory microculture. Journal of Personality and Social Psychology, 1971, 17, 179-191.

Whittaker, J. 0. The effects of experimentally introduced anchorages upon Judgments in the autokinetic situation. Unpublished doctoral dissertation, University of Oklahoma, 1958.

(Received June 27, 1975)


  1. This study was carried out as one of the research units of the Institute of Group Relations, University of Oklahoma with the support of the National Science Foundation. Further support was provided by National Institute of Mental Health Grant PHS MH 212-79/02.
    Thanks are given to Michael Lauderdale of the University of Texas who was of great assistance in conducting the experiment. Thanks are also due to Dorothy J. Pace, Center for Social Psychological Studies, Oklahoma State University, for suggestions on the manuscript; and to Thomas Haliburton, University of Texas, for providing the information in the illustrative case on the spelling of English. Richard Smith, Oklahoma State University, assisted in the analysis of the data.
    Requests for reprints should be sent to Mark K. MacNeil, who is now at the Department of Psychology, University of Stirling, Stirling FK9 4LA, Scotland.




Valid HTML 4.01 Strict Valid CSS2