Zajonc 1968 Attitudinal Effects of Mere Exposure

Download as pdf or txt
Download as pdf or txt
You are on page 1of 27

Journal of Personality and Social Psychology Monograph Supplement

Volume 9, No. 2, Partj2 June 1968

ATTITUDINAL EFFECTS OF MERE EXPOSURE 1


ROBERT B. ZAJONC University of Michigan The hypothesis is offered that mere repeated exposure of the individual to a stimulus object enhances his attitude toward it. By "mere" exposure is meant a condition making the stimulus accessible to the individual's perception. Support for the hypothesis consists of 4 types of evidence, presented and reviewed: (a) the correlation between affective connotation of words and word frequency; (b) the effect of experimentally manipulated frequency of exposure upon the affective connotation of nonsense words and symbols; (c) the correlation between word frequency and the attitude to their referents; (d) the effects of experimentally manipulated frequency of exposure on attitude. The relevance for the exposure-attitude hypothesis of the exploration theory and of the semantic satiation findings were examined.

exposure" is meant a condition which just makes the given stimulus accessible to the individual's perception. Even though the hypothesis seems to be in A mysterious student has been attending a class at Oregon State University for the past two months conflict with such celebrated laws as familienveloped in a big black bag. Only his bare feet arity breeds contempt and absence makes the show. Each Monday, Wednesday, and Friday at heart grow fonder, it is not particularly origi11:00 A.M. the Black Bag sits on a small table near nal or recent (Fechner, 1876, pp. 240-243; the back of the classroom. The class is Speech 113 basic persuasion. . . . Charles Goetzinger, professor James, 1890, p. 672; Maslow, 1937; Meyer, of the class, knows the identity of the person inside. 1903; Pepper, 1919). The foremost proponent None of the 20 students in the class do. Goetzinger of this hypothesis, the advertising industry, said the students' attitude changed from hostility has always attributed to exposure formidable toward the Black Bag to curiosity and finally to advertising potential. Butapparently, in refriendship [italics added]. spect for the law of enhancement by associaThis monograph examines the general hy- tionit seldom dared to utilize mere exposure. pothesis implied by the above phenomenon: The product, its name, or its hallmark is mere repeated exposure of the individual to a always presented to the public in contiguity stimulus is a sufficient condition for the en- with other and always attractive stimuli, comhancement of his attitude toward it. By "mere monly females, exposed more boldly than the 1 This work was supported by the National Insti- product itself. At the same time, however, the tute of Mental Health, Grant MH 12174. I wish to advertising industry also likes to warn against thank Christine Linder, Dik van Kreveld, and James owerexposure, relying, it would appear, on the J. Taylor for their invaluable assistance in carrying above law of familiarity (Erdelyi, 1940; out various phases of the experimental work. I am Wiebe, 1940). also indebted to Elinore Cottrell and Robert P. It isn't altogether clear just what evidence Weeks for making their students available as judges supports these advertising principles. And and subjects. On February 27, 1967, the Associated Press carried the following story from Corvallis, Oregon:
1968 by the American Psychological Association, Inc.

ROBERT B. ZAJONC direct evidence that attitudes are enhanced by mere exposure or mere contact with the stimulus object is scant. Moreover, it is the product of antiquated methods, and almost all of it concerns music appreciation (Downey & Knapp, 1927; Krugman, 1943; Meyer, 1903; Moore & Gilliland, 1924; Mull, 19S7; Verveer, Barry, & Bousfield, 1933; Washburn, Child, & Abel, 1927). The problem of attitudinal effects of social contact and interaction has also been of some interest in the study of interracial attitudes (Cook & Selltiz, 1952). But these studies have invariably examined the effects not of mere perceptual exposure of people to each other, but of processes considerably more complex: prolonged social interaction, group interdependence, cooperation, etc. (Deutsch & Collins, 1951; Kramer, 1950; MacKenzie, 1948; Wilner, Walkley, & Cook, 1952). Although the independent variables in these studies have generally been featured under the labels "contact" and "exposure," the effects they report cannot, because of confounding with a multitude of other events (and with reinforcement in particular), be regarded as produced alone by contact or exposure. Thus, it has been known for some time that social interaction enhances the attitudes of interactors toward each other (Bovard, 1951; Festinger, 1951; Homans, 1961; Newcomb, 1963). But it is not known just what contribution to the relationship between social interaction and attitudes is made by mere exposure on the one hand, and by the variety of psychologically significant processes that necessarily accompany mere exposure during the course of social interaction, on the other. The main empirical support for the exposure hypothesis comes, therefore, not from work on interaction, interracial attitudes, or attitudes in general, but from an entirely different and seemingly unrelated area of research. It comes from some recent work on word frequencies. This recent research shows that there exists an intimate relationship between word frequency and meaning. And this relationship, in my opinion (for which I shall later present support), may be a special case of the more general relationship between mere exposure and attitude enhancement. The strength and pervasiveness of the relationship between word frequency and meaningthe evaluative aspect of meaning, in particularis truly remarkable. For, if there is any correspondence between the frequency with which words are used and the actual preponderance of the things and events for which these words stand, then we may congratulate ourselves on living in a most happy world. According to the Thorndike-Lorge count (1944), the word "happiness" occurs 761 times, "unhappiness" occurs only 49 times. "Beauty" is to be found at least 41 times as often as "ugliness," and "wealth" outdoes "poverty" by a factor of 1.6. We "laugh" 2.4 times as often as we "cry"; we "love" almost 7 times more often than we "hate"; we are "in" at least 5 times more often than we are "out"; "up" twice as often as we are "down"; much more often "successful" than "unsuccessful"; and we "find" things 4.5 times more often than we "lose" themall because most of us are "lucky" (220) rather than "unlucky" (17). We have all the reasons in the world to be "happy" (1449) and "gay" (418) rather than "sad" (202) and "gloomy" (72), for things are 5 times more often "good" than "bad," almost 3 times more often "possible" than "impossible," and about five times more "profitable" than "unprofitable." That is, perhaps, why "boom" and "prosperity" outdo "recession" by a factor of just about 30, "abundance" outdoes "scarcity" by at least 3:1, and "affluence" is 6 times more prevalent than "deprivation." Catering to our corporeal sensibilities, things are 3 times more often "fragrant" than they are "foul," 12 times more often "fresh" than "stale," and almost 7 times more often "sweet" than "sour," and everything that can be filled is three times as often "full" as it is "empty." If we have anything, we have "more" of it 6 times more often than we have "less" of it,2 and 3 times more often "most" of it than "least" of it. And those things that we have so frequently more of are 5 times more often "better" than they are "worse," 6 times more often "best" than "worst," and 4 times more often "superior" than "inferior." Still,
2 N.B. The more-less ratio in this text is 7:1 up to now.

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE they "improve" at least 25 times as often as they "deteriorate." These examples suffice to convince one that the world represented by a one-to-one correspondence with word frequencies is as unreal as it is spectacular. Bitterly aware of it, Sartre (1964) confessed in his autobiography, " . . . as a result of discovering the world through language, for a long time, I took language for the world [p. 182]." But, while they are unfaithful in representing reality, word frequencies are extraordinarily accurate in representing real values: words that stand for good, desirable, and preferred aspects of reality are more frequently used. It isn't entirely clear who discovered this remarkable relationship between word frequency and the evaluative dimension of word meaning. Postman (1953) seems to be one of the early workers to note its generality, while Howes and Solomon (1950) observed in their critique of McGinnies' (1949) perceptual defense experiment that the so-called "taboo" words he used as stimuli are particularly infrequent. However, the first systematic research effort that demonstrates the wordfrequency-word-value relationship is due to Johnson, Thomson, and Frincke (1960). These authors were the first, I believe, to collect empirical data showing that words with "positive" meaning have higher frequency counts than words with "negative" meanings. They have also gathered experimental evidence showing that the repeated use of a nonsense word tends to enhance its rating on the good-bad scale of the semantic differential. Johnson, Thomson, and Frincke (1960) have not tried to explain either of these two aspects of the frequency-value relationship, being primarily concerned with its implications for the study of word-recognition thresholds. This paper examines the frequency-value relationship, proposing that it is considerably more pervasive and general than implied by the Johnson-Thomson-Frincke results, and that it is, moreover, a special case of a broader and more basic phenomenon: the enhancement of attitudes by mere repeated exposure. I shall first review evidence on the correlation between word frequency and word value, and between stimulus frequency and attitude. Experimental evidence on these two relationships, and on the likely causal direction, will then be examined. Word Frequency-Word Value: Correlational Evidence Johnson, Thomson, and Frincke (1960) obtained correlations of .63, .40, and .38 between the L-count (Thorndike & Lorge, 1944) and the good-bad scale values for three samples of randomly chosen words. In a further attempt, they constructed 30 pairs, each consisting of one frequent and one infrequent word. These pairs were given to a group of subjects with the instructions to "encircle the most pleasantly toned word of each pair." In 87% of the pairs the majority of subjects endorsed the more frequent word. Finally, 64 nonsense syllables of low, medium, and high association were rated by a group of subjects on the good-bad scale of the semantic differential. Johnson, Thomson, and Frincke reported a clear relationship between association value and "goodness" ratings. The rationale of this study invoked the assumed relationship between association of the given nonsense syllable and the probability of occurrence of the corresponding letter combination in meaningful words (Underwood, 1959). In an attempt to examine the generality of this phenomenon, we studied the evaluations of 154 antonym pairs. First, a large pool of antonym pairs was amassed. From this pool all symmetric8 pairs were chosen in the following manner. For each antonym pair 10 judges, 1 at a time, were asked to give the antonym of one member of the pair. Ten other judgesindependently of the first 10 were asked to give the antonym of the other
3 One finds in the course of this endeavor that the antonymic relation is seldom symmetric. According to the standard sources, if Y is listed as the antonym of X, then chances are that not X but Z is listed as the antonym of V. For instance, in the 1960 edition of Webster's New Collegiate Dictionary, "extend" is given as the antonym of "contract." Looking up "extend" we find, however, that its antonym is "reduce." The antonym of "reduce," on the other hand, is "increase." The antonym of "increase" is "decrease," the antonym of "decrease" is "amplify," the antonym of "amplify" is "condense," and the antonym of "condense" is "expand." We can ultimately close the circle, because "contract," according to this source, is the antonym of "expand."

ROBERT B. ZAJONC
TABLE 1 SEMANTIC PREFERENCE AND FREQUENCY OF 154 ANTONYM PAIRS

% Preferred agree- alternative (a) ment


100 100 100 100 100 100 100 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 99 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 98 97 97 97 97 97 97 97 97 97 97 97 97 97 97 [97 96
able attentive better encourage friendly honest possible advance best clean comfortable favorable good grateful peace present pure responsible reward right smile tolerant victory add advantage agreeable capable desirable find fortunate forward friend high honorable kind legal life love mature moral pleasant polite reliable success valid voluntary adequate competent found important likely on patience patient patiently popular positive profitable promote remember satisfactory willingly above

Nonpref erred alternative (6)

Frequency of (a)

Frequency of (W

% Preferred agree- alternative (a) ment

Nonpreferred alternative (&)

Frequency of (a)

Frequency of (6)

197

unable inattentive worse discourage unfriendly dishonest impossible retreat worst dirty uncomfortable unfavorable bad ungrateful war absent impure irresponsible punishment wrong frown intolerant defeat subtract disadvantage disagreeable incapable undesirable lose unfortunate backward enemy low dishonorable unkind illegal death hate immature immoral unpleasant impolite unreliable failure invalid involuntary inadequate incompetent lost unimportant unlikely off impatience impatient impatiently unpolular negative unprofitable demote forget unsatisfactory unwillingly below

930 49 2354 205 357 393


1289

452
1850

5122

781 348 93 194 472

239 4 450 147 19 41 459 105 292 221 112 25


1001 1118

13

197 267 154 3874


2143

1075

42 118
2018

404 58 176 160 2698 136 736 2553


1674 1521

65 4 30 80 890 216 13 166 6 41 43 30 42 593 108 139 883


1224

58

180 4804
5129

91 272 457 115 78 573 22 28 95 69 2892


1130

1074

8 34 34 815 756 17 19 114 3 9 262 56 26 59 23

364 30224 139 392 85 418 92 57 90


1682

154 66 941

40 25 3644 39 79 82 12 28 12 2 882 32 13 529

96 96 96 96 96 96 96 96 96 96 96 96 96 96 95 95 95 95 95 95 95 95 95 94 94 94 94 94 93 93 93 93 93 93 93 93 93 92 92 92 92 92 92 91 91 91 91 90 90 90 90 90 89 88 88 86 85 85 84 83 83 79 78

active early front full live presence probable rational reasonable resolutely strong succeed superior timely accept direct include increase most practical regularly rich wealth approve conscious leader obedient together agreement certain first major normal regular unselfish upwards wide more now up upward visible yes always familiar maximum optimism agree necessary over sweet whole light deep smooth white in independent fast comedy fasten day dry

passive late back empty die absence improbable irrational unreasonable irresolutely weak fail inferior untimely reject indirect exclude decrease least impractical irregularly poor poverty disapprove unconscious follower disobedient apart disagreement uncertain last minor abnormal irregular selfish downwards narrow less then down downward invisible no never unfamiliar minimum pessimism disagree unnecessary under sour part dark shallow rough black out dependent slow tragedy unfasten night wet

1022 1094 1129 4307

186

277 64 33 155 30 770 264 166 27 667 416 533 781 3443 340 122 656 243 171 299 373 70
1835

143 800
5154

366 335 340


32

8015 7665 11718

9 593

1663 2387 881 346 2663 75253 134 514 126 142 4549 592

111 110 2202 3285 345 43 28 729 715 7520 679

29 2859 6587 395 1079 163 14 9 56 4 276 620 40 6 51 23 38 86 1259 12 5 857 146 45 116 45 4 276 21 107 3517 83 43 44 137 40 391 1357 10208 5534 27 74 11742 5715 39 86 11 38 107 2961 102 1585 1005 104 294 1083 13649 18 434 189 16 3385 319

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE


TABLE IContinued
Preferred agree- alternative (a) ment

Nonpreferred alternative (i)

Frequency of (a)

Frequency of (6)
887

Preferred agree- alternative (a) ment


CO

Nonpreferred alternative (6)

Frequency of (a)

Frequency of (6) 1302

78 78 77 74 72 72 70 70 68 68 67 67 65 64

long unshaken usually upstairs inner interior near unlimited inside wrap infinite internal coming informal

short shaken unusually downstairs outer exterior

5362

far

6 718 314 143 185

1338

1835

83 91 226 97 48 67 921 17 2 26

limited outside unwrap finite external going formal

43 656 293 36

71

1486

64

4623

166

63 63 61 59 59 58 58 57 55 55 55 54 52 51

question women different same inward outward man women husband wife usual unusual offense defense hot cold import expert inwardly outwardly inconspicuous conspicuous work play mortal immortal answer

men

2132 3614 1194

43

2552 1747

54

7355 1788

516 86

2431 1668 1092

273 223

1006

86 32 33 54

88 33 59 26

2606

2720

member of the pair. Only those pairs were which there is relatively little agreement. For retained about which the 20 judges showed pairs with agreement greater than 95% (i.e., unanimous agreement with the dictionary the upper half of the list) there are only six sources. A list of 154 antonym pairs was thus reversals out of the 77 possible. It is signifiobtained. These were given to 100 subjects, cant, moreover, that in three of these six all college students, for judgments as to which antonym pairs the less desirable member member had "the more favorable meaning, (which in these cases is the more frequent represented the more desirable object, event, one) has more meanings and linguistic uses state of affairs, characteristic, etc." A differ- than the more desirable one. "Invalid" means ent random order of the antonym pairs was both "not valid" and "cripple," but "valid" given to each subject, and the lateral positions is just "valid." "Yes" is an adverb, but "no" of the members of each pair were reversed at is an adverb and an adjective. And "front" is a noun, a verb, and an adjective, while random for half of the group. Table 1 shows the list of these 154 antonym "back" is all that and an adverb to boot. Toward the end of the list where the desirpairs, together with the "desirability" and the frequency data (the Thorndike-Lorge L- ability preferences are divided fairly evenly count). The preferred member of each pair is between the two members of the antonym always listed first. The "desirability" figures pairs, the frequencies of the two antonyms are simply the percentages of subjects choos- often are nearly the same. "Play" is preferred ing the left member of the pair as the pre- to "work" only by a majority of two (a curious commentary on the contemporary college ferred alternative. It is of some interest, however incidental, population!), and the respective frequency that there is considerable agreement about counts of these antonyms are 2606 and 2720. desirability of the meanings. On half of the The "hot-cold" preference is 55 to 45 and items the agreement exceeded 95%. Agree- their frequency counts 1006 and 1092. The ment is high even for words which are not "husband-wife" preference is 58 to 42 and genuinely evaluative. For instance, 97 of the their respective frequencies, 1788 and 1668. Three antonym items about which agree100 students preferred "on" to "off," 98 preferred "add" to subtract, 96 "above" to ment was complete or nearly complete show a "below," and 92 "upward" to "downward." curious pattern of results. They are "goodFor the overwhelming majority of the items bad" (5122:1001), "better-worse" (2354: the preferred word is also the more frequent 450), and "best-worst" (1850:292). Since one. Only 28 of the 154 antonym pairs (18%) "better" is presumably better than "good," show a negative relationship between fre- "worse" worse than "bad," and since "best" quency and desirability. Moreover, these "re- is presumably better than "better," and versals" occur primarily for antonym pairs on "worst" worse than "worse," we would expect

ROBERT B. ZAJONC the greatest separation between the frequencies of "best" and "worst," smallest between the frequencies of "good" and "bad," and medium between the frequencies of "better" and "worse." Since absolute differences are deceiving, we best take the ratios of the frequencies, which are 6.34, 5.23, and 5.12 for "best-worst," "better-worse," and "good-bad," respectively. It is indeed the case that the frequency ratios increase from "good-bad" to "best-worse." However, if frequency reflects "desirability," we would also expect the frequency of "best" to exceed the frequency of "better," and that of "better" to exceed the frequency of "good." In fact, however, "good" is more frequent than "better," and "better" more frequent than "best!" But is "better" better than "good?" In an extensive study of meanings Hosier (1941) found that "good" was consistently rated as better than "better."
Startling as this may appear to grammarians, it is psychologically sound, since GOOD is a positive assertion, whereas BETTER implies comparison with some standard which might, in many cases, be itself unfavorable. Compare the often heard comment, "He is getting better, but he is still far from good" [p. 134].

For purposes of comparison the frequencies of French, German, and Spanish equivalents of some of the antonyms examined are given in Table 2 below. Systematic data on indigenous desirability ratings are unfortunately not available, but it would be surprising if the French, German, and Spanish judgments differed from those obtained in the United States. An informal inquiry among foreign visitors marshalled a good deal of support for this conjecture. Comparing the data in Tables 1 and 2, the agreement is rather striking. In 15 out of the 44 cases the frequency relation in the antonym pairs is the same in the three foreign languages as in English: the more favorable item is more frequent, a result exceeding chance expectation by a large margin. The results in Table 2, furthermore, give a ready expression to our favorite ethnic prejudices. The relatively low frequency of the two Romance equivalents of "early" and the high frequency of these equivalents of "late," in comparison to their Germanic counterparts, make generalizations about national character tempting, as does the relatively low frequency

of the German equivalent of "reward." The foreign equivalents of answer-question, hotcold, import-export, peace-war, etc., however, show patterns of differences that may reflect more than superficial linguistic idiosyncrasies. Several questions can immediately be raised about the above results. First, are these figures up to date? The Thorndike-Lorge count is based on samples of material published during the late twenties and the early thirties. The German equivalents come from a source dating to the late 19th century (Kading, 1898). The French count was published in 1929 (Van der Beke, 1929), and the Spanish in 1927 (Buchanan, 1927). Secondly, do these results reflect general verbal habits? Word counts are based on printed material alone. Do people show the same linguistic predilections in ordinary speech as they do in writing? Admittedly, both questions indicate caution in generalizing from the above results. But this caution needn't be excessive. Howes (1954) has recently asked Harvard and Antioch undergraduates to estimate the probabilities of various words. The correlations between the students' estimates of several word samples and the L-count of the Thorndike-Lorge source varied around .80. There is also evidence from word association studies showing that word counts do reflect general verbal habits of the population. A word which has a high frequency of occurrence in print is also a highly probable associate. The association norms to 200 words were recently collected by Palermo and Jenkins (1964) from a sample of 4,500 school children and college students in Minneapolis. The list of the 200 stimulus words represents a systematic sample of verbs, nouns, pronouns, adverbs, adjectives, participles, etc., all having fairly high frequency on the Thorndike-Lorge counts. Since in the word association task each subject makes one response to each stimulus word, Palermo and Jenkins collected from their subjects 900,000 word responses. Among them "good" occurred 4890 times, "bad" only 1956. The response "right" was given 477 times, the response "wrong" only 100 times. "Full" was found 431 times among the associations, "empty" only 62 times. "Strong" was given 557, "weak" 96 times. "Together" occurred 575 times, "apart"

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE TABLE 2 FREQUENCY RANKS OP ENGLISH, FRENCH, GERMAN, AND SPANISH ANTONYM PAIRS
English able (3) unable (9) accept (3) reject (9) active (6) passive (14) answer (2) question (3) better (2) worse (4) certain (2) uncertain (9) clean (3) dirty (7) comedy (9) tragedy (9) comfortable (5) uncomfortable (11) day (2) night (2) direct (3) indirect (12) dry (3) wet (4) early (2) late (2) fast (2) slow (3) find (2) lose (3) friend (2) enemy (3) full (2) empty (4) good (2) bad (2) high (2) low (2) hot (2) cold (2) husband (3) wife (3) import (7) export (11) increase (3) decrease (8) independent (6) dependent (14) life (2) death (2) light (2) dark (2) French capable (3) incapable (4) accepter (2) rejeter (5) actif (6) passif (?) reponse (4) question (2) meilleur (2) pire (S) certain (2) incertain (10) propre (2) sale (7) comedie (6) tragedie (9) a 1'aise (4) inconfortable (9) jour (2) nuit (2) direct (6) indirect (12) sec (3) mouil!6 (5) t6t (3) tard (2) vite (2) lent (4) trouver (2) perdre (2) ami (2) ennemi (2) plein (2) vide (4) bon (2) mauvais (2) haut (2) bas (2) chaud (3) froid (3) mari (3) femme (2) importation (11) exportation (10) augmentation (10) reduction (11) independent (7) ddpendent (?) vie (2) mort (2) clair (3) sombre (3) German fahig (4) unfahig (11) annehmen (2) ablehnen (5) tatig (S) untatig (?) Ant wort (3) Frage (2) besser (2) schlechter (?) sicher (2) unsicher (9) sauber (9) schmutzig (12) Komodie (9) tragodie (11) bequem (S) unbequem (10)
Tag (2) Nacht (2)

Spanish capaz (3) incapaz (7) acceptar (3) rechazar (S) activo (6) pasivo (10) respuesta (4) pregunta (4) mejor (2) peor (2) cierto (2) incierto (9) limpio (3) sucio (6) comedia (4) tragedia (8) coniodo (7) incomodo (10) dia (2) noche (2) directo (4) indirecto (8) seco (3) mojado (6) temprano (4) tarde (2) pronto (2) lento (4) encontrar (2) perder (2) amigo (2) enemigo (2) lleno (2) vacio (4) buen (2) mal (2) alto (2) bajo (2) caliente (5) frio (2) esposo (2) esposa (2) importaci6n (?) exportaci6n (13) aumento (5) diminuci6n (?) independiente (S) dependiente (9) vida (2) muerte (2) claro (2) obscuro (2)

direkt (3) indirekt (8) trocken (S) nass (9) frtih (2) spat (2) schnell (2) langsam (3) finden (2) verlieren (2) Freund (2) Feind (2) voll (2) leer (4) gut (2) schlecht (3) hoch (2) niedrig (4) heiss (S) kalt (3) Mann (2) Frau (2) Einfuhr (11) Ausfuhr (12) Vermehrung (6) Verminderung (11) selbststiindig (4) abhiingig (6) Leben (2) Tod (2) hell (4)

dunkel (3)

ROBERT B. ZAJONC
TABLE 2 Continued
English French German Spanish

life (2) die (2) long (2) short (2) love (2) hate (4) more (2) less (2) near (2) far (2) peace (3) war (2) positive (9) negative (11) possible (3) impossible (5) presence (4) absence (7) reward (6) punishment (6) right (2) wrong (3) strong (2) weak (3) sweet (2) sour (9) together (2) apart (4) victory (5) defeat (7) wealth (4) poverty (7) white (2) black (2) wide (2) narrow (3)

vivre (2) mourir (2) long (2) court (3) aimer (2) hair (6) plus (2) moins (2) pres (3) loin (2) paix (3) guerre (3) positif (6) negatif (11) possible (2) impossible (3) presence (2) absence (5) recompense (6) punition (12) juste (2) faux (3) fort (2) faible (3) doux (2) amer (4) ensemble (2) s6par6 (2) victoire (4) defaite (8) richesse (5) pauvrete (12) blanc (2) noir (2) large (2) etroit (3)

leben (2) sterben (2) lang (2) kurz (2) lieben (2) hassen (6) mehr (2) weniger (2) nah (2) fern (2) Friede (3) Krieg (2) positiv (8) negativ (?) moglich (2) unmoglich (3) Anwesenheit (9) Abwesenheit (9) Anerkennung (S) Strafe (4) richtig (2) falsch (3) stark (2) schwach (3) suss (4) sauer (9) zusammen (2) getrennt (3) Sieg (4) Niederlage (8) Vermogen (4) Armut (10) weiss (2) schwartz (3) breit (4) schmal (6)

vivir (2) morir (2) largo (2) corto (3) amar (2) odiar (7) mas (2) menos (2) cerca (2) lejos (2) paz (2) guerra (2) positive (7) negative (7) posible (2) imposible (2) presencia (3) ausencia (4) premio (4) castigo (4) justo (3) mal (2) fuerte (2) debil (4) dulce (2) amargo (4) junto (2) separado (3) victoria (5) derrota (9) riqueza (3) pobreza (S) bianco (2) negro (2) ancho (3) angosto (8)

Note.The figures in brackets indicate frequency ranks: (1) means that the word is among the 500 most frequent words, (2) that it is among the 1000 most frequent words, (3) that it is among the 1500 most frequent words, etc. The source of these counts is Eaton iton (1940).

29 times. "Light" was a response 8655 times (N.B., some subjects must have given it more than once), "dark" 4274 times. But as in the case of the Thorndike-Lorge count, "front" occurred 22 times, while "back" occurred 265 times; "rich" was given 36 times, while "poor" was a response 95 times. "Near" was given 981 times, "far" 1218. "Coming" was given 166 times, "going" 714 times. And, as in Lcount, "play" and "work" showed 791 and 957 occurrences, respectively. However, the best evidence about the relationship between the individual's verbal hab-

its and the evaluative aspect of meaning is found in a recent study by Siegel,* although it wasn't the purpose of her study to explore this relationship. Siegel's experiment dealt with the effects of verbal reinforcement on the emission of words differing in affective connotation and in frequency. Eighteen six-letter words of known frequencies and previously judged on the good-bad and the pleasantunpleasant scales were selected from a larger
* Siegel, Felicia S. Effects of word frequency and affective connotation on verbal responding during extinction. (Mimeo)

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE sample. Six of these words were of high frequency (100 and more in 1 million), six of medium (20 to 30), and six of low frequency (1 to 5). Within each frequency class two words were previously judged to be good, two neutral, and two bad. Three groups of subjects, other than those involved in the affective judgments, participated in the experiment, each having to deal with six words of the same frequency. The procedure consisted of presenting the subject with the list of six words, all high, medium, or low in frequency, depending on the condition in which he was in, and giving him at the same time a stack of cards on which appeared illegible six-letter "words." Ostensibly, each card contained one of the six words in the subjects' list. Actually, the "words" consisted of random sequences of six letters, printed over several thicknesses of paper and one carbon. Their legibility was further reduced by placing each card in an onionskin paper envelope. The subjects' task was to "read" or to guess what word appeared on each card. Of interest for the present purposes are the first SO trials which served to establish operant rate, and during which, of course, no reinforcement of any sort was given. Table 3 shows data on the guessing behavior of Siegel's subjects as a function of word frequency and affective connotation. Reported in each cell is the average number of times a word of a given frequency and affective value was used as a guess during the SO operant trials. Since there are six words to choose from, 8.33 represents a chance response rate. It is clear, however, that both frequency and affective connotation displace response rate away from the chance level. High frequency seems to result in overcalling, and low freTABLE 3
FREE RESPONSE EMISSION AS A FUNCTION or WORD-FREQUENCY AND WORD VALUE"
Word frequency
... . .

Low

Medium

High

Good Medium

Bad X

7.43 6.28 6.28 6.66

9.43 8.57 5.86 7.95

9.68 8.71 7.71 8.70

8.8S 7.85 6.61

' From Siegel, I960.

quency in undercalling. But it is striking to discover that affective connotation had an even stronger effect on response emission, the marginals for that variable showing a somewhat greater range of differences. Some words in the language have primarily an evaluative function. These words should show the frequency-value relationship with particular clarity. Several instances of this relationship are examined. Let us first consider the scales of the Semantic Differential (Osgood, Suci, Si Tannenbaum, 1957). We have chosen only those scales which have high and relatively pure loadings on one of the three main factors, evaluation, potency, and activity. Table 4 shows the polar opposites of these scales, together with their frequencies according to the Thorndike-Lorge L-count. The left-hand polar opposites in the three columns are the favorable, potent, and active ends of the scales. It is significant that among the 19 evaluative scales the favorable polar opposite has always higher frequency than the unfavorable opposite. For the scales which do not load high on the evaluative factor the high frequencies are divided fairly evenly among the potent and nonpotent opposites. In 9 of the IS potency scales the highly potent end of the scale is more frequent. In 3 of the 8 activity scales the active polar opposite is more frequent. There are two other instances of a high correlation between frequency and value for adjectives. The first comes from the work by Gough (1953). Gough has given the items of his Adjective Checklist to 30 judges who rated each adjective for favorability. The most favorable and the least favorable quartiles of Gough's checklist are reported in his publication. The average word frequency of the upper quartile is 140, and of the lower quartile 48. The second illustration comes from data collected by Anderson (1964). A list of SSS adjectives was recently used by Anderson in his work on impression formation. The list was constructed out of a large sample of items. The 555 selected items were given by Anderson to a group of 100 subjects with the instructions to rate on a 7-point scale "how much you yourself would like the person described by that word." We have simply computed the correlation between these likeabil-

10

ROBERT B. ZAJONC
REQUENCY PER 4,SOO,OOO WORDS (L-COUNT 01 01 01 01 O O O O O 0 0 O O

TABLE 4
POLAR OPPOSITES OF THE SEMANTIC DIFFERENTIAL AND THEIR FREQUENCIES Evaluative factor beautiful clean fair fragrant good grateful happy harmonious honest kind nice pleasant positive reputable sacred successful sweet true wise ugly dirty unfair foul bad ungrateful

987 781 561 66


5122 1449

178 221 59 39
1001

.>
<

^s
r- 83

194

sad

dissonant dishonest cruel awful unpleasant negative disreputable profane unsuccessful sour false foolish Potency factor

26 393 630 457 92 23 102 352 679 420

1521

13 202 9 41 165 370 114 28


21

/X^
I.OI2.013.OI- 4.01- 5.OI2.OO 3.OO 4.OO 5OO 6.OO AVERAGE FAVORABILITY RATING

o-

1711

13 14 102 209 223

I.OO

FIG. 1. Average frequencies of 555 adjectives rated for favorability. (Based on data from Anderson, 1964.)

frequencies are plotted for six categories of adjectives in increasing order of favorability. Considering that the reliabilities of the Thorn1909 dike-Lorge count and of Anderson's favor1005 680 ability ratings are less than perfect, the co1697 1818 efficient of correlation of .83 is particularly 40 54 99 91 impressive. 294 346 Miller, Newman, and Friedman (19S8) 248 37 9 119 have shown that word frequency is a negative 770 276 function of word length. The problem im22 7 mediately arises, therefore, as to which of 646 443 391 593 these two variables is critical for word value and word meaning. In order to examine this Activity factor possible confounding between frequency and word length, the above correlation was recom434 514 active passive puted holding the number of letters constant. 1005 bright dark 645 267 excitable calm 7 No appreciable change in the previously ob514 434 slow fast tained coefficient was observed. 21 2 heretical orthodox The relationship between word frequency 1092 1006 hot cold 48 rash 37 cautious and word length is generally explained in 289 sharp 324 dull terms of the principle of least effort. Words that require considerable effort in writing and in speech are less likely candidates for use. ity ratings and the logarithm of the Thorn5 In an attempt to control for effort Frincke dike-Lorge L-count. Figure 1 shows this and Johnson (1960) have asked subjects to relationship graphically, where means of log choose the "most pleasantly toned word" from 5 Items for which there was no frequency informa- each of 108 homophone pairs. The greatest tion in the Thorndike-Lorge count were not included majority of these pairs consisted of words of in computing this coefficient. These items were prithe same length, and all pairs, of course, conmarily of the hyphenated form, such as open-minded, good-humored, well-spoken, fault-finding, ultra-criti- sisted of words that required the same effort in uttering them. Out of 3,132 possible choices, cal, wishy-washy, etc.
bass brave deep hard heavy large masculine mature rough rugged severe strong tenacious thick wide treble cowardly shallow soft light small feminine youthful smooth delicate lenient weak yielding thin narrow

28 216 881

17 26 104 549

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE the more frequent member of the pair was chosen 1,836 times. Dixon and Dixon (1964) have given a list of 200 verbs (in past-tense form) to 60 female and 60 male judges who rated them on an 11-point good-bad scale. The instructions were to rate what "kind of impression the subject thought a psychologist would get of him when he used each verb in a sentence." These impression ratings have correlations with log frequencies (the Thorndike-Lorge Lcount) equal to .48 for females and to .50 for males. But it must be pointed out that these coefficients represent correlations severely attenuated by unreliability of the frequency variable. The Thorndike-Lorge count lists verbs in the present-tense form. If an adjectival form of the verb exists, then it is also listed. In our own research, in computing correlation coefficients, only the present-tense frequencies were used. Miron (1961) had American and Japanese subjects rate a sample of three-element phonetic combinations on various scales of the Semantic Differential. The subjects also rated these stimulus materials for their familiarity. It is interesting that the correlations between familiarity and the composite of evaluative scales were .59 and .50 for the American and the Japanese samples, respectively. But the correlations of familiarity with the composites of the potency and activity factors were low and negative. As a final example of the relationship between word frequency and the evaluative aspect of meaning, two poems by William Blake are called to the reader's attention: Infant Joy "I have no name: I am but two days old," What shall I call thee? "I happy am, Joy is my name." Sweet joy befall thee! Pretty joy! Sweet joy but two days old, Sweet joy I call thee: Thou dost smile, I sing the while, Sweet joy befall thee! Infant Sorrow My mother groaned! My father wept; Into the dangerous world I leapt; Helpless, naked, piping loud, Like a fiend hid in a cloud, Struggling in my father's hands, Striving against my swadling bands, Bound and weary I thought best To sulk upon my mother's breast.

11

In these two poems, expressing opposite qualities of affect, the frequencies of the critical words (i.e., words which convey the major content, and hence not articles, pronouns, or auxiliary verbs) were averaged. The average frequency of Infant Joy is 2,037. The average for Infant Sorrow is 1,116. Two formally similar verses, one by Browning and the other by Shelley, show the same pattern: Song. R. Browning The year's at the spring, And day's at the morn; Morning's at seven; The hillside's dew-pearled; The lark's on the wing; The snail's on the thorn; God's in his Heaven All's right with the world. Dirge. P. B. Shelley Rough wind, that meanest loud Grief too sad for song; Wild wind, when sullen cloud Knells all the night long; Sad storm, whose tears are in vain, Bare woods, whose branches strain, Deep caves and dreary main Wail, for the world's wrong. The average word frequency of Browning's poem is 1,380. The poem by Shelleywhich comes to a rather different and sadder conclusionhas an average frequency of 728. Stimulus Frequency Attitude: Correlational Evidence We may now turn to the more general question of the effect of exposure on attitude, still limiting ourselves to correlational studies. Here, less evidence exists, and the evidence which is available is often indirect. But the

12

ROBERT B. ZAJONC Lorge L-count, and whose frequencies can be arranged in increasing order in approximately constant log units. These countries were then given to high-school students with the instructions to rank-order them in terms of liking. Table 5 shows the average rank each country received and its frequency of usage according to the L-count. There seems to be little question about the frequency-attitude relationship. The same relationship is found with American cities. Selected were 10 cities that (a) are listed in the Thorndike-Lorge L-count, and (b) can be arranged in increasing order of frequency in approximately constant log units. University students were asked how much they would like to live in each of these 10 cities. Their task, specifically, was to rankorder these cities according to their preferences "as a place to live." The average ranks, together with frequency counts of these 10 cities, are shown in Table 5. Other subjects, also high-school students in the Midwest, were asked to rate on a 7point scale how much they liked various trees, fruits, vegetables, and flowers. In each case 10 items were selected which were listed in the Thorndike-Lorge count and which could be ordered according to a constant log frequency unit. Table 6 shows both the average ratings (0 = dislike; 6 = like) and the frequency counts for the four types of items. The rank correlations between the frequency and average attitude are .89, .85, .84, .81, .85, and .89, for countries, cities, trees, fruits, vegetables, and flowers, respectively. TABLE 5
PREFERENCE RANKS AND FREQUENCY COUNTS FOR 10 COUNTRIES AND 10 CITIES
Countries AverFre- age prefquency erence rank Cities AverFre- age prefquency erence rank

results are quite similar to those just reviewed. For instance, Alluisi and Adams (1962) found a correlation of .843 between the preference subjects expressed for the appearance of letters and their frequency in the language. Strassburger and Wertheimer (1959) had subjects rate for "pleasantness" nonsense syllables varying in association value. Higher association values consistently received higher "pleasantness" ratings. Wilson and Becknell (1961) and Braun (1962) successfully replicated these results. Braun also found that eight-letter pseudo-words, varying in their order of approximation to English (Miller, 1951), show the same pattern. These two studies differ from the similar ones by Johnson, Thomson, and Frincke, discussed earlier, in that subjects in the former ones were asked to judge how pleasant were the stimuli themselves, or how much subjects liked them (Wilson & Becknell, 1961), while in the latter whether they meant something close to "good" or close to "bad." In 1947 the National Opinion Research Center conducted an extensive survey on the "prestige" of various occupations and professions. Nearly 100 occupational categories were rated for "general standing." Twentyfour of these occupations are labeled by single words, such as "physician," "scientist," "janitor," etc. The remainder is described less economically: "owner-operator of a printing shop," or "tenant farmerone who owns livestock and machinery and manages the farm." Thus, one is able to determine the frequency of usage for only a part of this listthe 24 single-word occupations. The correlation between rated occupational prestige of these 24 items and the log of frequency of usage is .55. Similar to the ratings of occupational prestige are the social distance ratings of ethnic and racial groups, first developed by Bogardus (1925) over 30 years ago. Recent replications show that these social distance ratings enjoy remarkable stability (Bogardus, 1959). The correlation between the so-called "racial-distance quotients," which are numerical equivalents of these ratings, and the log frequency of usage of these ethnic labels is .33. In order to explore relationships of this sort further, I have selected 10 countries whose names are found in the Thorndike-

Country

City

England Canada Holland Greece Germany Argentina Venezuela Bulgaria Honduras Syria

497 130 59 31 224 15 9 3 1


4

2.67 3.33 3.42 4.00 4.92 6.08 6.58 7.75 7.92 8.34

Boston Chicago Milwaukee San Diego Dayton Baltimore Omaha Tampa El Paso Saginaw

255 621 124 9 14 68 28 5 1 2

2.75 3.08 3.83 4.25 5.75 6.08 7.08 7.08 7.50 7.58

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE


TABLE 6 PREFERENCE RATINGS or TREES, FRUITS, VEGETABLES, AND FLOWERS, AND THEIR CORRESPONDING FREQUENCIES
Trees

13

f 172 75 125 8 34 14 2 1 3 4

APR

Fruits apple cherry strawberry pear grapefruit cantaloupe avocado pomegranate gooseberry mango

APR

Vegetables

APR

Flowers rose lily violet geranium daisy hyacinth yucca woodbine anemone cowslip

APR

pine walnut

oak fir

rosewood birch sassafras aloes

yew
acacia

4.79 4.42 4.00 3.96 3.83 3.75 3.00 2.92 2.83 2.75

220 167 121 62 33 1.5 16 8 5 2

5.13 5.00 4.83 4.38 4.00 3.75 2.71 2.63 2.63 2.38

corn potato lettuce carrot radish asparagus cauliflower broccoli leek parsnip

227 384 142 96 43 5 27 18 3 8

4.17 4.13 4.00 3.57 3.13 2.33 1.96 1.96 1.96 1.92

801 164 109 27 62 16 1 4 8 2

5.55 4.79 4.58 3.83 3.79 3.08 2.88 2.87 2.54 2.54

Note.f = frequency of usage; APR = average preference rating.

Of course, word counts do not faithfully represent the frequencies with which one encounters the above items. And it is difficult to discover precisely how often the average Midwestern high school student encounters a yew, a cowslip, or a radish. But a fair index of frequency of exposure can be found in farm production data. For seven of the vegetables in Table 6 farm production figures for 1963 are available, and they are shown below in thousands of tons: corn (4.17) potatoes (4.13) lettuce (4.00) carrots (3.57) asparagus (2.33) cauliflower (1.96) broccoli (1.96) 2,340.9 13,777.1 1,937.6 843.8 187.8 123.4 123.9

question of which is the cause and which the effect remains a matter of conjecture. We shall now turn, therefore, to such experimental evidence. ExposureMeaning: Experimental Evidence
Experiment I. The first experimental study on the relationship between exposure and word meaning was carried out by Johnson, Thomson, and Frincke (1960). These authors first asked subjects to rate a number of nonsense words on the good-bad scale of the semantic differential. The subjects were then instructed that "this is an experiment concerning the effectiveness of repetition in learning to pronounce strange words correctly." Some of these words were shown once, others twice, 5 times, or 10 tunes. Subjects were required to look at these words and to pronounce them on each presentation. Following this training procedure the words were again rated on the good-bad scale. A significant exposure effect was obtained, with the words shown frequently increasing on the evaluative scale. Strangely, however, words which were seen only once in training were judged afterwards not quite as "good" as before training. Thus, as a result of 2, 5, and 10 exposures words improved in meaning, and as a result of but 1 exposure they deteriorated. This finding, however, may be an artifact of the beforeafter procedure used by Johnson, Thomson, and Frincke. Moreover, frequencies and stimuli were fully confounded in their study. Our experiment used the same stimuli which, incidentally, came from the familiar experiment by Solomon and Postman (19S2) on the effects of word frequency on recognition threshold, but our design differed from the one used by Johnson, Thomson, and Frincke in several respects. In the JohnsonThomson-Frincke experiment the same words always appeared in the same frequencies to all subjects. Thus, the word "jandara," for instance, was given 10 times to each subject, and the word "mecburi" was given once to each subject. It is possible that the effects these authors obtained are not due to the

Included also (in brackets) are average preference ratings of these seven vegetables. The rank correlation between the production figures and the average preference ratings is .96. Of course, this impressive correlation coefficient, like those we observed above, may not reflect the effect of frequency on attitude but the effect of attitude on frequency. Thus, it can be argued that many roses are grown because people like roses. But it can also be argued that people like roses because there are many roses growing. There is less ambiguity, however, with regard to the correlation between frequency of letters and the preference for their appearance (Alluisi & Adams, 1962). There aren't so many e's in English just because we like the way e's look. Still, until there is experimental evidence, the

14

ROBERT B. ZAJONC
rately. The ratings of "goodness" were averaged for each word when it was given during training with the lower frequencies of 0, 1, and 2 (hatched bars), and when it was given with the higher frequencies of S, 10, and 25 (solid bars). This was possible because each word was used in each frequency equally often but for different subjects. It is evident from Figure 3 that some words are rated as having more positive meaning than others, and this effect is indeed significant (F = 8.35; # = 11/781; p<.001). Apparently, some of these words "sound better" than others. But independently of word content, subjects consistently rated the given word to mean something "better" if they had seen it (and had said it) more often. This is true for all the 12 words used in the experiment, a result that has a chance likelihood equal to .00024. Experiment II. Since the hypothesis proposed above holds that it is mere exposure that is a sufficient condition of attitude change, the procedure used in Experiment I is not optimal for testing its validity. Subjects in Experiment I were required to pronounce the nonsense words during training, and it is possible that a decrease in difficulty in pronouncing the words associated with successive presentations was responsible for the results. In other words, subjects rated the frequent stimuli more favorably because they found them easier to pronounce than stimuli which they saw and pronounced only once or twice. And, there were stimuli which they never pronounced and, in fact, did not really know how to pronounce. Wilson and Becknell (1961) suggested that the evaluative ratings of nonsense syllables of high association value are higher than of low associ-

frequency manipulation alone, but that they depend on the stimulus material with which the frequency variable was fully confounded. In our study words and training frequencies were, therefore, counterbalanced in a Latin-square design. Because words and the number of exposures were counterbalanced, an after-only design could be employed, requiring no premeasures. The effects of repeated exposure could be observed for each word by comparing the favorability rating it received after having been exposed during training once, twice, five times, etc. Eliminating premeasures also eliminated for each stimulus one full exposure that necessarily preceded and therefore accompanied the frequency manipulation. The present experiment differs from that of Johnson, Thomson, and Frincke (1960) in several other respects which are less critical for the interpretation of results. The procedure of this experiment, therefore, is described in some detail. Except for some specific changes, the same general methodology is followed throughout this series of studies. Twelve seven-letter "Turkish" words, shown in Figure 3, were counterbalanced against six frequencies (0, 1, 2, 5, 10, and 25) in six replications of the experiment. Seventy-two subjects were run, one at a time, 12 subjects in each replication. The initial instructions informed the subject that the experiment dealt with "pronouncing foreign words." He was told that he would be shown some foreign words, hear the experimenter pronounce them, and that he would be required to pronounce them himself. The words were typed on 3 X S-inch cards. On each trial a card was shown to the subject for approximately 2 seconds. Simultaneously the experimenter pronounced the word, requiring the subject to follow him. Since each frequency class contained two word-stimuli, there were 86 trials altogether. The position of a given stimulus in the sequence of these 86 trials was determined at random. Following the above frequency training subjects were told that the words they had just learned to pronounce were in fact Turkish adjectives, and that their next task would be to guess what they meant. The experimenter told the subject that he realized how nearly impossible this task was, and he therefore did not require him to guess the word meanings exactly. Instead, it would suffice if the subject indicated on a 7-point (0 to 6) good-bad scale whether each word meant something good or something bad and to what extent, because these Turkish adjectives all meant something good or bad. These ratings were made of the 10 stimuli which the subject received during the frequency training, and of 2 additional ones previously never seen by him. The results of the experiment are shown in Figure 2 and in Figure 3. In Figure 2 are shown the ratings of "goodness" averaged for each of the six frequencies, and plotted on a log scale. Each point on that curve is based on 144 observations, and it is clear that a strong exposure effect was obtained (F = 5.64; d/= 5/355; /><.001). Figure 3 shows the exposure effect for each of the 12 words sepa-

CHINESE CHARACTERS NONSENSE WORDS

0 1 ]

FREQUENCY

FIG. 2. Average rated affective connotation of nonsense words and Chinese-like characters as a function of frequency of exposure.

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE

IS

sufficient factor in affecting evaluative ratings but not a necessary one. In order to eliminate the pronounciation factor and to reduce the subjects' active LOW FREQUENCY participation while exposure is being manipulated, AFWORBU the following experiment was carried out. HIGH FREQUENCY To meet the requirements of the definition of "mere exposure," Chinese characters were substituted for the nonsense words. These stimuli were taken from Hull's (1920) concept formation study, and I am told that not only are most of them meaningless, but that they are also far from the absolutely NANSOMA minimal standards of Chinese caligraphy. Nevertheless, they were quite adequate for our experimental purposes. The subjects were again told that the experiment dealt with the learning of a foreign language, but now they were not required to proENANWAL. nounce the characters. Nor were they able to pronounce them subvocally. They were simply instructed to pay close attention to the characters whenever they were exposed to them. In all other ZABULON respects the experiment was identical to the one employing nonsense words. Now, too, following training subjects were told that the characters stood LOKANTA for adjectives, and that their task was to guess their meaning on the good-bad scale. Characters and exposures were again counterbalanced. Figures 2 and 4 show the results, and it is obvious that the exposurefavorability relationship previously found with nonsense words obtains (F = 4.72; dj = 5/335; p < .001) even if the individual's exposure to the stimu1 2 3 4 5 lus consists of his passively looking at it for a RATED "GOODNESS" OF MEANING period of about 2 seconds. Figure 4 shows that the FIG. 3. Average rated affective connotation of exposure effect is found for all stimuli but one. The above results add strength to the hypothesis nonsense words exposed with low and high frequenthat mere exposure is a sufficient condition for atticies. tude enhancement. But again the last experiment did ation value because they are easier to pronounce. In not succeed in completely eliminating a learning facorder to follow up their suggestion a group of 22 tor from the exposure manipulation, for it is possiUniversity of Michigan subjects were given the ble that this manipulation is now confounded with above "Turkish adjectives" with the instructions to the ease of recognition. This danger of confounding, rate them according to "how easy or difficult it is to however, is probably minimal because at no time pronounce" them. Using a 7-point scale, a signifi- were the subjects ever required to recognize or cant item-effect was revealed by an analysis of vari- discriminate the idiograms. The results of Experiments I and II are in an ance (F= 14.28; d f = 11/263; p<.00l), showing that there are indeed differences among the nonsense apparent conflict with results reported by Amster words in the ease with which they can be pro- and Glasman (1966). These researchers report a nounced upon their first presentation. The Wilson- negative result using a procedure similar to that Becknell conjecture is supported by a correlation of employed by Johnson, Thomson, and Frincke (1960). .46 between the average ease of pronouncing the The experiment was similar in all respects except words and the evaluative scores obtained in Experi- that meaningful English words were substituted for ment I. These latter scores were obtained by the nonsense stimuli. No exposure effect was obaveraging for each word the rating it obtained in all served by Amster and Glasman for these meaningful words. But this finding is not at all surprising. Nor frequencies. These results, however, in themselves do not is it especially significant for the understanding of preclude a relationship between exposure and evalua- exposure effects. Adding one more occasion (or even tive rating. With ease of pronouncing held constant 10 more occasions) to see and say a perfectly wellthe exposure effect may still be obtained. This ex- known English word to all the times this word had pectation is strengthened by the results of a study been seen and uttered by the individual in the past described above (Frincke & Johnson, 1960) in which a figure often in the thousandsreally shouldn't have homophone pairs differing in word-frequency were much effect on the meaning he attributes to it. The rated for "pleasantness." Since homophones do not expectation of a change in the evaluative aspect of differ in pronounciation, the obtained frequency meaning as a function of a few additional exposures effects show that ease of pronouncing may be a becomes even less reasonable when we consider that

16

ROBERT B. ZAJONC use. Because the second item of evidence rests on experimental proof, in which the frequency of usage was systematically and independently manipulated, one cannot question the causal direction implied in these data. But finding that the frequency of usage affects meaning needn't necessarily preclude the possibility that meaning determines the frequency of usage. It is necessary, therefore, to examine more closely the results on the correlational evidence between word-frequency and word value. Why are positive words used more frequently? Besides the rather wistful and unlikely explanation that there are more positive than negative referents (i.e., we live in a paradise), one real possibility suggests itself. The evidence reviewed so far deals only with usage per word. The totality of "good" and "bad" usage, however, depends on the numbers of different "good" and "bad" words in the language. It is entirely possible, therefore, that the superiority of "good" words in frequency per word exists side by side with the superiority of "bad" words in their greater variety. This possibility receives some support from the fact that in English (and in a host of other languages) prefixes and suffixes that serve to negate or reverse meaning, such as anti, de, im, in, ir, less, un, etc., are most commonly attached to words having a positive connotation. Once attached to a word they almost universally form a word with a negative affective connotation. Positive words with these prefixes or suffixes are exceptional: unselfish, independent, are some examples. It would appear, therefore, that there are indeed more negative than positive words. And if there are more different negative words, the usage per word would naturally be attenuated for these words, because the total usage would be distributed among a larger universe of items. Norman has asked a group of students to separate a large sample of adjectives into "good" ones and "bad" ones. On the average 2.31 more items were placed in the "bad" pile than in the "good" pile. The frequency figures in Table 1 above show a pattern consistent with Norman's independent finding.
Warren T. Norman, personal 1965.
0

LOW FREQUENCY HIGH FREQUENCY

RATED "GOODNESS" OF MEANING

FIG. 4. Average rated affective connotation of Chinese-like characters exposed with low and high frequencies. the change in affective connotation is a linear function of the logarithm of frequency, as we noted in Figures 1 and 2. If n is the frequency of the subject's preexperimental exposure to the word, then the comparisons made by Amster and Glasman involved the following four frequencies: n + 1, n-\-2, n + S, and n + 10. Since n is large, perhaps as large as 1000, the differences in exposures amounted to fractions of 1%.

Word-Frequency-Word-Value Relationship as a Special Case oj the Exposure-Attitude Relationship In the first section of this paper some evidence was presented suggesting that words with positive affective connotations are used more frequently (both in print and in speech) than words with negative affective connotations. In the second section evidence was given to suggest that the affective connotation of a word improves with their repeated

communication,

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE The average frequency of the preferred antonyms is 2.3 times larger than the average frequency of the nonpreferred antonyms! Therefore, for the material considered here, the ratio of total positive and negative usage is equal to unity. If repeated usage enhances the affective meaning of words, a relatively large supply of negative words would in fact be needed. It would be equally reasonable to expect that there exist devices in language protecting words from a deterioration of meaning. It is entirely possible that the prefixes and suffixes discussed above serve this function. Because the negative qualities of these prefixes and suffixes are independent of their referents, because they are essentially abstract, and because they derive their negativity from the semantic function they perform, words formed by means of these prefixes and suffixes are perhaps better able than root words to resist an enhancement of affective connotation as a result of repeated usage. I was unable to find evidence corroborating this point of view, although there is a good deal of philological literature on both positive changes in meaning (see for instance vanDongen, 1933) and negative changes in meaning (see, for instance, Schreuder, 1929). Most of the sources, however, consider changes in meaning of root words only. If there are many remaining doubts that frequency of words is a function of the value of their referents, then the following frequencies of a few well-chosen but significant words should once and for all dispel them: Psychologist Chemist Economist Sociologist Astronomer Geologist Physicist Geographer Botanist Biologist

17

36 32 32 14 12 9 8 7 6
S

Exposure-Attitude Relationship: Experimental Evidence


Experiment HI. In all the experiments above the question asked of the subjects in rating the stimuli

following exposure dealt with the evaluative aspect of their meaning. The subjects were never required to say just how much they "liked" the nonsense words or "Chinese" characters. In all probability, the results would have been the same if they were asked directly to state their attitude toward these words and characters, and the Wilson-Becknell (1961) results support this conjecture. But because their stimuli were essentially verbal, subjects' answers could in these studies be strongly influenced by semantic factors. This would have been less likely, of course, in the case of Chinese characters than in the case of nonsense words. As was pointed out above, there is some direct evidence on the attitudinal effects of mere exposure, dealing almost exclusively with music appreciation. Meyer (1Q03), for example, played to his students oriental music 12 to IS times in succession. In most cases the students' introspective protocols indicated a better liking for the pieces on the last than on the first presentation. One of the students who took part in Meyers' experiment (H. T. Moore), and who showed enhancement effects of repeated exposure ("I liked the last time better than the first, because I became more used to the successive chords"), followed up this work in a study of his own 20 years later. Moore and Gilliland (1924) played to their students jazz and classical records once a week for 25 weeks. Liking for classical records increased, but no change was found for jazz music. Similar results are reported by other writers (Krugman, 1943; Verveer, Barry, & Bousfield, 1933; Washburn, Child, & Abel, 1927). Downey and Knapp (1927) played to 33 students a variety of musical selections (e.g., Tschaikowsky's Marche Slave, Massenet's Meditation from "Thais," Columbia, The Gem of the Ocean, etc.) once a week for five weeks. All pieces of music except one (Columbia, The Gem of the Ocean) became better liked at the close of the sessions. Alpert (1953) presented subjects with sounds having unfamiliar rhythms. His subjects found these sounds at first unpleasant. After repeated presentations, however, the liking for them increased. Additional exposures of subjects to the tones resulted in increasing indifference on the part of the listeners. More recently, Mull (1957) found that upon repeated exposure to their music subjects enjoyed Schoenberg and Hindemith more. In the area of visual arts, Pepper (1919) found that repeated exposure resulted in more positive esthetic judgments of unusual color combinations. Krugman and Hartley (1960), however, using famous paintings, could only find ambiguous results. Maslow (1937) projected for 4 days in succession 15 paintings of great masters. Six days following the last presentation the 15 paintings were presented once again, and interspersed among them were 15 others (matched for the artist) which the subjects had never seen. The results indicated a greater liking for the familiar paintings. Maslow (1937) also made tests of preference, frequently with similar results, for other familiar and unfamiliar objects, such as rubber bands, paper clips,

18

ROBERT B. ZAJONC

blotters, pens, pencils, etc. A similar experiment to the one with paintings, but using instead Russian girls' names, showed the same results. The same subjects were used in all these studies and the sessions took place in the same room, the subjects always sitting in the same chairs. Toward the end of the testing program Maslow asked if anyone would like to change seats. No one did, preferring, apparently, to remain in the familiar one. Although the results of the above studies are fairly consistent, the conditions under which they were carried out make their conclusions somewhat less than compelling. In the majority of instances, the circumstances of the repeated exposure were quite ambiguous. The experiments were usually conducted in classes, the instructor serving as the experimenter. Subjects often responded aloud, thus being able to influence each other's judgments and opinions. Prior to the sessions the experimenter often expressed his own preferences. The stimuli, repeatedly shown, were not always exposed under the same conditions, and the material, exposures, and sequences were seldom counterbalanced. But in all of these experiments a pattern of results emerges showing that the frequency manipulation has more pronounced attitude effects for stimuli that are novel, unfamiliar, or unusual than for familiar stimuli. This pattern is, of course, consistent with the observation that attitude enhancement is a function of the logarithm of frequency. Becknell, Wilson, and Baird (1963) have recently reported more convincing support for the exposureattitude hypothesis. Slides of nonsense syllables were presented with different frequencies (1, 4, 7, and 10). Following this exposure training (which also included interspersed presentations of slides with landscapes and with ads) female subjects were given pairs of boxes containing nylon stockings, and they were asked to choose the "brand" they preferred. These "brands" corresponded to the nonsense syllables previously shown, and they were printed on the boxes. Each subject received two different pairs of boxes for comparison. The paired-comparison data showed a tendency of subjects to prefer the box marked by the more frequent syllable. Again, however, the semantic component is not excluded from the effects obtained in these two studies. There is one more item of evidence, somewhat indirect, on the problem of the effects of exposure. In a study by Munsinger (1964) subjects were given the opportunity to present to themselves CVC trigrams whose association value, evaluation scale value, and prepotency score (Handler, 19SS) were previously assessed. By pressing a response key the subject would expose in a small window a trigram which he would then have to spell. The rate at which he key-pressed constituted the dependent measure. In one of Munsinger's experimental groups subjects could expose to themselves, by means of that key response, trigrams that were matched for association and prepotency. All these trigrams, however, previously scored low on the evaluative scales of the semantic differential. After subjects reached

t 3.5

J3.0

2.5 -

2
FREQUENCY

10

25

FIG. 5, Average attitude toward photographs as a function of frequency of exposure. an asymptotic key-pressing rate, the experimental conditions changed such that now the subjects' response would expose trigrams that were high in evaluation, although they were still matched for association and prepotency. A significant increase in key-pressing rates is reported by Munsinger following the change in the affective value of the trigrams. Again, however, the semantic component is not entirely excluded from the effects obtained in these two studies. Because they are less a matter of semantic factors, we have chosen to manipulate interpersonal attitudes by means of exposure. Using the same experimental design as with the Chinese characters, faces of men (photographs of graduating Michigan State University seniors taken from the MSU Yearbook) were employed as attitude objects. The experiment was introduced to subjectsall students at the University of Michiganas dealing with the problem of "visual memory." Following the exposure manipulation, which consisted of presenting each photograph a different number of times for a period of 2 seconds, subjects were asked to rate on a 7-point scale how much, they might like the man on each photograph. The results of this study are shown in Figures S and 6. While the exposure effect is not as clear as previously (only 9 of the 12 stimuli show it), it is still rather impressive (F = 9.96; dl = 5/ 355; p< .001).

The Exposure-attitude Hypothesis and Related Theoretical Issues The above results raise a series of empirical and theoretical questions. Are all attitudes

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE

19

*-~

&*

*% & $

Vfti x

i * 4*

arfL

?^

^ \SfA

FIG. 6. Average attitude toward photographs exposed with low and high frequencies.

enhanced by mere repeated exposure? Is there a number of repetitions beyond which attitude begins to become negative? Does this number vary systematically across attitude objects? Are these effects stable? These and similar questions can only be answered by further empirical work. On a theoretical level these questions address themselves primarily to those psychological processes that mediate exposure effects. Let us first consider a possible biological significance of an exposure-enhancement mechanism. A stimulus presented for the first time evokes in the organism an instinctive fear reaction. Lorenz (1956) noted that a young raven,
confronted with a new object, which may be a camera, an old bottle, a stuffed polecat, or anything else, first reacts with escape responses. He will fly up to an elevated perch, and, from this point of vantage, stare at the object literally for hours. After this he will begin to approach the object very gradually . . .

Biihler, Hetzer, and Mabel (1928) observed that human infants reacted to a strange sound by crying out with fear. Upon the second exposure of the sound stimulus, movement and vocalization that indicated displeasure were observed. On the third exposure, the infants listened to the sound showing some signs of attention, but did not seem to show any displeasure. On the fourth exposure, they looked in the direction of the sound with detectable

interest. These facts, of course, are borne out by common observation. Hunt (1965) reported that young infants he observed preferred a familiar mobile to a new one. And the "Black Bag" story cited in the introduction represents another example of phenomena in this category. At the outset the "Black Bag," in fact, attracted a good deal of hostility. Cairns (1966) has recently presented a very convincing argument that the affiliative behavior and social attachments among animals are solely determined by the animals' exposure to one another. Examining evidence on affiliative preferences of animals observed under conditions of inter- and intraspecific cohabitation and of animals deprived of social contact, Cairns concluded that such affiliative preferences vary directly with the length of the association and with the importance of the cues which are generated in the course of the association. Cairns, moreover, did not limit his conclusion to inter-animal social attachments but prposed that "animals tend to remain in the presence of [any] objects to which they have been continually exposed [p. 409]." The survival value of an avoidance reflex to a novel stimulus is obvious. But there is no direct evidence that all organisms are equipped with an avoidance reaction occurring upon the encounter of a novel stimulus. However, if we assume that they are, then the exposure-attitude hypothesis becomes more

20

ROBERT B. ZAJONC
stimulus were examined in an independent experiment. Fifteen subjects were presented with nonsense words (the same as in Experiment I) in a series of 86 trials. Two words appeared 25 times, two 10 times, two 5 times, two twice, and two once. The position of a word in the series of trials was determined by a random device. Words were counterbalanced against frequencies in three experimental replications. Due to a mechanical failure of the apparatus, data for one subject could not be used, and were not included in the analysis. On a given trial, the stimulus word was projected onto a screen for a period of 2 seconds, and the subject's GSR was recorded. The interstimulus interval was 20 seconds. GSR was measured only during the first 10 seconds following stimulus onset, and all GSRs occurring during the last 12 seconds of the interstimulus interval were treated as artifacts. The Kaplan-Hobart technique of GSR measurement (Kaplan & Hobart, 1964), which requires current of only 10 microamperes, was employed. Zinc-zinc sulphate electrodes (Kaplan & Fisher, 1964; Lykken, 1959) were applied to the forefinger and the middle finger of the subject's nonpreferred hand. The results of the experiment are shown in Figure 7 and Figure 8. In Figure 7 changes in conductance are plotted for each successive presentation of the stimulus-word. For purposes of clarity data for only one stimulus are graphed in each frequency class. The results for the other set of stimuli are the same. It can be seen that, in general, successive presentations result in a lower autonomic reactivity. After about seven or eight exposures a stable asymptote is reached. Hence, only stimuli shown 25 and 10 times attain an asymptote. Words shown five times, twice, or once generate greater changes in conductance even on their last exposure. This effect is seen better in Figure 8 in which GSRs on the last exposure of the stimuli were plotted. As prior frequency of exposure increases there is a

reasonable. The first encounter with the novel stimulus produces fear reaction. If no negative consequences are associated with this first encounter, the avoidance reaction upon the second encounter will naturally be weaker. If such encounters continue, and if no other eventsnegative in their consequences for the organismaccompany these encounters, then the organism's attitude toward the stimulus must improve. To be sure, the hypothesis does not deny or preclude the effects of reinforcement. The exposure of a stimulus coupled with reward will strengthen the animal's approach behavior; and the exposure of stimulus coupled with a noxious event will strengthen his avoidance reactions. But in the absence of reward or punishment, mere exposure will result in the enhancement of the organism's attitude toward the given stimulus object. If novel stimuli evoke fear, conflict, or uncertainty, one should be able to detect these states, and to observe their dissipation upon repeated exposure. To the extent that GSR measures arousal that is associated with the above states, we would expect greater GSRs upon the presentation of novel stimuli than upon the presentation of familiar ones, and we would also expect a drop in GSR reactivity to be the consequence of repeated stimulus exposure.
Experiment IV. Changes in affective arousal that occur as a result of repeated exposure of a novel

.35
30

.25
20 .15
.10

25 Exposures . .o 10 Exposures 5 Exposures ..7 Z Exposures I Exposures

5.05
Llj

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

FIG. 7. GSR obtained upon repeated exposures of nonsense words.

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE


40

21

2 5 10 FREQUENCY OF PAST EXPOSURE (LOG SCALE)

2 5

FIG. 8. GSR obtained on the last exposure of nonsense words exposed with various frequencies. lesser change in conductance upon stimulus presentation. This effect is significant at the .OS level (F = 4.01; df = 4/117). These results cannot be due to an overall adaptation that may be occurring as the series of the 86 trials progresses. It is clear that the stimulus shown just once generated a substantial GSR although it occurred as late as on the 67th trial. As a matter of fact, its GSR does not differ from the GSR obtained for the first presentation of the other stimuli. The last presentation of the stimulus word that had just one prior exposure also resulted in a substantial GSR although it occurred on the 64th trial. And on the whole, although there is variability over trials, at any one point during the series GSR's are higher for stimuli with infrequent prior exposures than for stimuli previously seen by the subject many times. It seems, therefore, that with increased exposure there is a genuine reduction in stimulus-evoked arousal. These findings are in agreement with those reported by Berlyne, Craw, Salapatek, and Lewis (1963). In their study subjects were shown visual patterns differing in complexity and in incongruity. The patterns were presented for 3 seconds on three successive occasions. A significant drop in GSR was obtained between the first and the second presentation. But no significant GSR effects were associated with complexity or incongruity.

DISCUSSION While the bulk of the results presented and reviewed in this monograph supports the hypothesis that repeated exposure is a sufficient condition of attitude enhancement, there are findings and theoretical formulations which

appear to be in conflict with the hypothesis. The most pronounced source of ostensibly contradictory results is in the area of exploration and curiosity. There is impressive evidence today that in a free situation the subject (human or animal) will turn toward a novel stimulus in preference to a familiar one (e.g., Berlyne, 1960). If such orienting and exploratory "approach" behavior is a symptom of a favorable attitude toward the stimulus object, the wealth of data on exploration and orienting behavior (Berlyne, 1950, 19SS; Berlyne & Slater, 1957; Dember &Milbrook, 1956; Montgomery, 1953; Thiessen & McGaugh, 1958; Thomson & Solomon, 1954) stands in clear contradiction to those reported in this monograph. But there is at present no direct evidence to support the above assumption. And, on the contrary, it is more likely that orienting toward a novel stimulus in preference to a familiar one may indicate that it is less liked rather than it is better liked. Ordinarily, when confronted with a novel stimulus the animal's orienting response enables it to discover if the novel stimulus constitutes a source of danger. It need not explore familiar stimuli in this respect. Novelty is thus commonly associated with uncertainty and with conflictstates that are more likely to produce negative than positive affect. Most recent work by Harrison (1967) indicates quite clearly that exploration and favorable attitudes are in fact negatively related. Using nonsense words, Chinese characters, and photographs of men's faces, Harrison obtained measures of liking from one group of subjects and measures of exploration from another group. The correlations between exploration and liking were .69, .69, and .60 for nonsense words, characters, and photographs, respectively. If the function of orienting behavior is eventually to change the novel stimulus into a familiar one, it is also its consequence to render the stimulus object eventually more attractive (or perhaps merely less repulsive). In his research Harrison also obtained data on the behavioral consequences of exposure, and in particular, on response conflict that novelty seems to arouse. It is a truism, of

22

ROBERT B. ZAJONC (e.g., bad) end of the scale, but are rated toward the neutral point of the scale. While several studies have demonstrated a reduction of polarization following rapid repetition of a word (Das, 1964; Kanungo & Lambert, 1963a, 1963b; Messer, Jakobovitz, Kanungo, & Lambert, 1964), there is an equal amount of conflicting evidence (Amster & Glasman, 1965; Floyd, 1962; Reynierse & Barch, 1963; Schulz, Weaver, & Radtke, 1965). Yelen and Schulz (1963) attribute satiation findings to a regression artifact. A reduction of polarity of positive words as a result of repetition would indeed be embarrassing for the exposure-attitude hypothesis. A reduction of polarity of negative words (i.e., words with a semantic score below the neutral point) would, of course, be entirely in agreement with the present results, for they would simply be showing an enhancement effect along the evaluative dimension. It should be noted that all our stimuli initially received negative ratings, that is, below the neutral point, 3, on the 0-6 good-bad scale. Given no exposure at all, the evaluative ratings for the Turkish nonsense words, Chinese-like characters, and photographs were 2.56, 2.67, and 2.79, respectively. In terms of polarity these averages are .44, .33, and .21. Contrary to the semantic satiation hypothesis, absolute polarities increased after 25 exposures, for they were +.61, +.78, and +.61, respectively. Admittedly, the controversy within semantic satiation literature may have to be resolved before clear implications for the exposure effect can be drawn. Parenthetically it should be noted, however, that the form of exposure used in the semantic satiation paradigm (i.e., 30 to 45 repetitions in 15 seconds) is not what has been above defined as the sufficient condition of attitude enhancement. Moreover, in semantic satiation studies the stimulus is commonly a verbal response made by the subject himself. Whether such a response-produced stimulus should constitute what is meant by mere exposure is a matter of some doubt. The above hypothesis and data seem to be consistent with the theory of reinforcement recently proposed by Premack (1959). This new and engaging approach to reinforcement effects, for which a good deal of impressive

course, to assert that novel stimulus is one to which no specific response, beside orienting, has as yet been attached. But the novel stimulus cannot fail being similar to an entire host of other stimuli that the individual had encountered in the past, and to which he had attached specific responses. And it is entirely likely that some of these generalized response tendencies that the novel stimulus simultaneously excites are mutually incompatible; that is, they cannot all be emitted at the same time. This antecedent condition is what makes response conflict (of some, however small, magnitude) a necessary concomitant of novelty. Using latency of free associations as a measure of response conflict, Harrison was able to demonstrate that response conflict is markedly reduced upon repeated exposure of a novel stimulus. Chinese ideographs were shown to a group of subjects once, twice, S times, 10 times, and 25 times, in a manner similar to that used in other experiments described above. Following this exposure manipulation, Harrison presented each stimulus once again requiring the subject to respond "with the first thing that came to mind." Included now were also Chinese ideographs that the subject had never seen. The latency of these free associations was obtained on each such trial, and the results revealed a systematic and significant drop in latencies as a function of the frequency of prior exposure. Another set of data which may also be of some consequence for the exposure-attitude hypothesis is to be found in the area of semantic satiation. In a typical semantic-satiation experiment the subject is asked to repeat words, two or three per second, for a period of IS seconds. The general findings in this area indicate that following this sort of rapid repetition the word seems to "lose" its meaning (for a review of the literature see Amster, 1964). Loss of meaning is measured by a departure from polarity on semantic differential scales, such as good-bad, strongweak, etc. (Lambert & Jakobovits, 1960). When repeated in rapid succession and rated on some semantic differential scale, immediately thereafter the words tend to be placed neither toward one (e.g., good) nor the other

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE evidence has already been accumulated (Premack, 1961, 1962, 1963, 1965; Premack & Bahwell, 1959; Premack & Collier, 1962), holds that if the emission of one response, A, is made contingent upon the emission of another response, B, and if A occurs with greater frequency than B, B will gain in the rate of emission. A can, therefore, be considered as having positive reinforcement value which seems to depend alone upon its frequency of occurrence. Contingencies between responses can be introduced by controlling manipulanda or the availability of goal objects. Thus, for instance, in studying these effects in Cebus monkeys, Premack (1963) used four manipulanda: a plunger, a hinged door, a vertically operated lever, and a horizontally operated lever. It was possible to make any manipulandum inoperable at any time, and to make it operable only in the case of a prior manipulation of another manipulandum. Thus, each manipulandum could serve for the reinforcwg or for the reinforced response. After establishing for each monkey the probabilities of operating each item under free access to all, contingencies were arranged between pairs of items, such that, for instance, the vertical lever could not be operated unless the animal pulled the plunger, or the hinged door remained locked unless the subject pressed the horizontal lever, etc. The introduction of these contingencies resulted in the predicted effects. In general, responses that were less probable increased in the rate of emission when more probable responses were made contingent upon them. When two responses were equal in probability of emission little or no change in response rates was observed. Viewed in the present context, the individual's response probabilities can be taken as an indication of his "attitudinal liking" or "attraction" to the goal objects of these responses, or to the instrumental stimuli associated with them. To the extent that the reward value of a given response, A, which is contingent upon another response, B, is a direct function of A's probability of emission, then increasing its probability of emission will increase its reward value. Premack (1961) was able to demonstrate such an effect in rats

23

for licking and bar pressing. Changing the language somewhat, it may be said that the individual's "attitudinal liking" for the goal object of the response A increases with the individual's exposure to the goal object of A. In Premack's work continuous reinforcement schedules are used and, hence, the individual's frequency of exposure to the goal object of A is equal to the frequency of emission of A. There are important differences between the phenomena observed by Premack and those reported in the present paper. In most of his experiments the frequency of occurrence of responses results from an independent, perhaps genetically given or previously acquired, response preference of the subject. In the experimental work on attitudinal effects of exposure, the frequency of occurrence of the stimulus-object is deliberately manipulated. Moreover, the subject usually does not have a prior preference for the stimulus exposed. On the contrary, he may often manifest avoidance tendencies. But the parallel is compelling. While there is a great deal in common between Premack's work on reinforcement and the research on attitudinal effects of exposure, a clear understanding of the implications of one for the other requires a systematic determination of what "reinforcement value" and "attitudes" have in common that makes them both equally vulnerable to simple frequency effects. CONCLUSION The balance of the experimental results reviewed and reported in this paper is in favor of the hypothesis that mere repeated exposure of an individual to a stimulus object enhances his attitude toward it. But, as yet, the account books cannot be closed. Further research must examine the boundary conditions of the exposure-attitude relationship, for it is possible that the neat linear log-frequency-attitude relationship, repeatedly observed here, may well break down under some conditions. This future research must, in particular, concentrate on the effects of large frequencies of exposure, on duration of exposure, on interexposure intervals, and on

24

ROBERT B. ZAJONC tion of training greater numbers of correct responses were associated with smaller percentages of reinforcement. However, in the various conditions of reinforcement subjects worked until they performed in succession a predetermined number of correct responses, the same for all percentages of reinforcement. Therefore, percentage of reinforcement was in this study completely confounded with the number of stimulus exposures (and also with the number of reinforcements). Although many authors have tried to cope with this confounding in one way or another (e.g., Festinger, 1961; Kanfer, 1954; O'Connell, 196S), the methodological difficulties have not been completely overcome. None of the four variables that are associated with the partial reinforcement effect-percentage of reinforced trials, number of trials, number of positive reinforcements, number of nonreinforcements can be studied independently of the others.
REFERENCES AIXUISI, E. A., & ADAMS, O. S. Predicting letter preferences: Aesthetics and filtering in man. Perceptual and Motor Skills, 1962, 14, 123-131. ALPERT, R. Perceptual determinants of affect. Unpublished master's thesis, Wesleyan University, 1953. AMSTER, H. Semantic satiation and generation: Learning? Adaptation? Psychological Bulletin, 1964, 62, 273-286. AMSTER, H., & GLASMAN, L. D. Verbal repetition and connotative change. Journal of Experimental Psychology, 1966, 71, 389-395. ANDERSON, N. H. Likeableness ratings of 555 personality-trait adjectives. Los Angeles: University of California, 1964. (Mimeo) BECKNEIX, J. C., JR., WILSON, W. R., & BAIRD, J. C. The effect of frequency of presentation on the choice of nonsense syllables. Journal of Psychology, 1963, 56, 165-170. BERLYNE, D. E. Novelty and curiosity as determinants of exploratory behavior. British Journal of Psychology, 1950, 41, 68-80. BERLYNE, D. E. The arousal and satiation of perceptual curiosity in the rat. Journal of Comparative and Physiological Psychology, 1955, 48, 232246. BERLYNE, D. E. Conflict, curiosity, and exploratory behavior. New York: McGraw-Hill, 1960. BERLYNE, D. E., CRAW, M. A., SALAPATEK, P. H., & LEWIS, J. L. Novelty, complexity, incongruity, extrinsic motivation, and the GSR. Journal of Experimental Psychology, 1963, 66, 560-567. BERLYNE, D. E., & SLATER, J. Perceptual curiosity, exploratory behavior, and maze learning. Journal

many other similar parameters of mere exposure. This research must also assess the applicability of the exposure-attitude relationship to a greater variety of stimulus objects. The question of generalization of specific exposure effects is of equal theoretical importance. Does repeated exposure to a given stimulus result in the enhancement of attitudes toward similar or related stimuli? Because the above effects seem to be a function of the logarithm of frequency of exposure, they are more apparent and more pronounced for differences among small frequencies than for differences among large frequencies. For the same reason, attitudinal enhancement produced by means of exposure will be more readily effected for novel objects than for familiar ones. It is likely that exposure effects for very familiar objects are absent completely or are so small that they cannot be detected at all by methods now available. As we have seen above, Amster and Glasman (1966) failed to obtain the exposure effect for common English words. It will be important for future research, therefore, to determine the range of familiarity for which the exposure effect is obtained. Mere exposure is a necessary precondition of a vast variety of experimental manipulations. For example, in attempts to change attitudes by means of persuasive communications the attitude object is mentioned repeatedly, regardless of whether the attempt is directed toward making the attitude more favorable or toward making it less favorable. Making attitudes more favorable should, therefore, be easier than making them less favorable. It is interesting that studies on the effectiveness of persuasion in attitude change seldom try to effect a negative change, and almost never compare the relative success of a pro-persuasion with the success of a conpersuasion. In an attitude-change study Tannenbaum and Gengel (1966) have recently obtained only positive shifts, although both a positive as well as a negative manipulation were employed. The partial reinforcement manipulation, too, is subject to possible confounding with the number of stimulus presentations. Erlebacher and Archer (1961), for instance, reported the curious result that at the comple-

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE


of Comparative and Physiological Psychology, 1957, 50, 228-232. BOGAEDUS, E. S. Measuring social distance. Journal of Applied Sociology, 1925, 9, 299-308. BOGARDUS, E. S. Social distance. Yellow Springs, Ohio: Antioch Press, 1959. BOVAED, E, W., JR. Group structure and perception. Journal of Abnormal and Social Psychology, 1951, 46, 398-405. BRAUN, J. R. Three tests of the McClelland discrepancy hypothesis. Psychological Reports, 1962, 10, 271-274. BUCHANAN, M. A. Graded Spanish word book. Toronto: Toronto University Press, 1927. BIJHLER, C., HETZER, H., & MABEL, F. Die Affektwirksamkeit von Fremheitseindrticken im ersten Lebensjahr. Zeitschrift fur Psychologie, 1928, 107, 30-49. CAIRNS, R. B. Attachment behavior of mammals. Psychological Review, 1966, 73, 409-426. COOK, S. W., & SELLTIZ, C. Contact and intergroup attitudes: Some theoretical considerations. New York: Research Center for Human Relations, 1952. DAS, J. P. Hypnosis, verbal satiation, vigilance, and personality factors: A correlational study. Journal of Abnormal and Social Psychology, 1964, 68, 7278. DEMBER, W. N., & MILLBROOK, B. A. Free choice by the rat of the greater of two brightness changes. Psychological Reports, 1956, 2, 465-467. DEUTSCH, M., & COLLINS, M. E. Interracial housing: A psychological evaluation of a social experiment. Minneapolis: University of Minnesota Press, 1951. DIXON, T. R., & DIXON, J. F. The impression value of verbs. Journal of Verbal Learning and Verbal Behavior, 1964, 3, 161-165. DOWNEY, J. E., & KNAPP, G. E. The effect on a musical programme of familiarity and of sequence of selections. In M. Schoen (Ed.), The effects of music. New York: Harcourt, Brace, 1927. EATON, H. S. An English, French, German, Spanish word frequency dictionary. New York: Dover, 1940. ERDELYI, M. The relation between "Radio Plugs" and sheet sales of popular music. Journal of Applied Psychology, 1940, 24, 696-702. ERLEBACHER, A., & ARCHER, E. J. Perseveration as a function of degree of learning and percentage of reinforcement in card sorting. Journal of Experimental Psychology, 1961, 62, 510-517. FECHNER, G. T. Vorschule der Aesthetik. Leipzig: Breitkopf & Hartel, 1876. FESTINGER, L. Group attraction and membership. Journal of Social Issues, 1951, 7, 152-163. FESTINGER, L. The psychological effects of insufficient reward. American Psychologist, 1961, 16, 1-11. FLOYD, R. L. Semantic satiation: Replication and test of further implications. Psychological Reports, 1962, 11, 274. FRINCKE, G., & JOHNSON, R. C. Word value and word frequency in homophone pairs. Psychological Reports, 1960, 7, 470,

25

GOUGH, H. G. Reference handbook for the Gough Adjective Check List. Berkeley: University of California Institute of Personality Assessment and Research, April 1955. (Mimeo) HARRISON, A. A. Response competition and attitude change as a function of repeated stimulus exposure. Unpublished doctoral dissertation, University of Michigan, 1967. ROMANS, G. C. Social behavior: Its elementary forms. New York: Harcourt, Brace, 1961. HOWES, D. On the interpretation of word frequency as a variable affecting speed of recognition. Journal of Experimental Psychology, 1954, 48, 106-112. HOWES, D. H., & SOLOMON, R. L. A note on McGinnies' "Emotionality and perceptual defense." Psychological Review, 1950, 57, 229-234. HULL, C. L. Quantitative aspects of the evolution of concepts. Psychological Monographs, 1920, 28(1, Whole No. 123). HUNT, J. McV. Traditional personality theory in the light of recent evidence. American Scientist, 1965, 53, 80-96. JAMES, W. The principles of psychology. Vol. 2. New York: Holt, 1890. JOHNSON, R. C., THOMSON, C. W., & FRINCKE, G. Word values, word frequency, and visual duration thresholds. Psychological Review, 1960, 67, 332-342. KADING, F. W. Haufigkeitsworterbuch der deutschen Sprache. Berlin: Mittler, 1948. KANFER, F. H. The effect of partial reinforcement in acquisition and extinction of a class of verbal responses. Journal of Experimental Psychology, 1954, 48, 424-432. KANUNGO, R. N., & LAMBERT, W. E. Paired-associate learning as a function of stimulus and response satiation. British Journal of Psychology, 1963, 54, 135-144 (a). KANUNGO, R. N., & LAMBERT, W. E. Semantic satiation and meaningfulness. American Journal of Psychology, 1963, 76, 421-428 (b). KAPLAN, S., & FISHER, G. R. A modified design for the Lykken zinc electrodes. Psychophysiology, 1964, 1, 88-89. KAPLAN, S., & HOBART, J. L. A versatile device for the measurement of skin resistance in rats and humans. American Journal of Psychology, 1964, 77, 309-310. KRAMER, M. Residential contact as a determinant of attitudes toward Negroes. Unpublished doctoral dissertation, University of Chicago, 1950. KRUOMAN, H. E. Affective response to music as a function of familiarity. Journal of Abnormal and Social Psychology, 1943, 38, 388-392. KRUGMAN, H. E., & HARTLEY, E. C. The learning of tastes. Public Opinion Quarterly, 1960, 24, 621631. LAMBERT, W. E., & JAKOBOVITS, L. A. Verbal satiation and changes in the intensity of meaning. Journal of Experimental Psychology, 1960, 60, 376-383. LORENZ, K. L'instinct dans le comportement de I'animal et de I'homme. Paris: Masson, 1956.

26

ROBERT B. ZAJONC
PEPPER, S. C. Changes of appreciation for color combinations. Psychological Review, 1919, 26, 389-396. POSTMAN, L. The experimental analysis of motivational factors in perception. In J. S. Brown (Ed.), Current theory and research in motivation. Lincoln: University of Nebraska Press, 1953. PREMACK, D. Toward empirical behavior laws: I. Positive reinforcement. Psychological Review, 1959, 66, 219-233. PREMACK, D. Predicting instrumental performance from the independent rate of the contingent response. Journal oj Experimental Psychology, 1961, 61, 163-171. PREMACK, D. Reversibility of the reinforcement relation. Science, 1962, 136, 255-257. PREMACK, D. Rate differential reinforcement in monkey manipulation. Journal of the Experimental Analysis oj Behavior, 1963, 6, 81-89. PREMACK, D. Reinforcement theory. Nebraska Symposium on motivation, 1965, 13, 123-180. PREMACK, D., & BAI-IWELL, R. Operant-level lever pressing by a monkey as a function of interest interval. Journal of the Experimental Analysis of Behavior, 1959, 2, 127-131. PREMACK, D., & COLLIER, G. Joint effects of stimulus deprivation and intersession interval: Analysis of nonreinforcement variables affecting response probability. Psychological Monographs, 1962, 76(5, Whole No. 524). REYNIERSE, J. H., & BARCH, A. M. Semantic satiation and generalization. Psychological Reports, 1963, 13, 790. SARTRE, J. P. Words. New York: Braziller, 1964. SCHREUDER, H. Pejorative sense development. Groningen: Noordhoff, 1929. SCHULZ, R. W., WEAVER, G. E., & RADTKE, R. C. Verbal satiation?? Psychonomic Science, 1965, 2, 43-44. SOLOMON, R. L., & POSTMAN, L. Usage as a determinant of visual duration thresholds of words. Journal of Experimental Psychology, 1952, 43, 195201. STRASSBURGER, F., & WERTHEIMER, M. The discrepancy hypothesis of affect and association value of nonsense syllables. Psychological Reports, 1959, 5, 528. TANNENBAUM, P. H., & GENOEL, R. W. Generalization of attitude change through congruity principle relationships. Journal of Personality and Social Psychology, 1966, 3, 299-304. THIESSEN, D. D., & McGAUGH, J. L. Conflict and curiosity in the rat. Paper presented at the meeting of the Western Psychological Association, Monterey, California, 1958. THOMSON, W. R., & SOLOMON, L. M. Spontaneous pattern discrimination in the rat. Journal of Comparative and Physiological Psychology, 1954, 47, 104-107. THORNDIKE, E. L., & LORGE, I. The teacher's wordbook of 30,000 words. New York: Teachers College, Columbia University, 1944.

LYKKEN, D. T. Properties of electrodes used in electrodermal measurements. Journal of Comparative and Physiological Psychology, 19S9, 52, 629634. MACKENZIE, B. K. The importance of contact in determining attitudes toward Negroes. Journal of Abnormal and Social Psychology, 1948, 43, 4. HANDLER, G. Associative frequency and associative prepotency as measures of response to nonsense syllables. American Journal of Psychology, 19SS, 68, 662-66S. MASLOW, A. H. The influence of familiarization on preference. Journal of Experimental Psychology, 1937, 21, 162-180. MC&NNIES, E. Emotionality and perceptual defense. Psychological Review, 1949, 56, 244-2S1. MESSER, S., JAKOBOVITS, L. A., KANUNGO, R. N., & LAMBERT, W. A. Semantic satiation of words and numbers. British Journal of Psychology, 1964, 55, 156-163. MEYER, M. Experimental studies in the psychology of music. American Journal oj Psychology, 1903, 14, 456-476. MILLER, G. A., NEWMAN, E. B., & FRIEDMAN, E. A. Length-frequency statistics for written English. Information and Control, 1958, 1, 370-398. MIRON, M. S. A cross-linguistic investigation of phonetic symbolism. Journal of Abnormal and Social Psychology, 1961, 62, 623-630. MONTGOMERY, K. C. Exploratory behavior as a function of "similarity" of stimulus situations. Journal of Comparative and Physiological Psychology, 1953, 46, 129-133. MOORE, H. T., & GILLILAND, A. R. The immediate and long time effects of classical and popular phonograph selections. Journal of Applied Psychology, 1924, 8, 309-323. MOSIER, C. I. A psychometric study of meaning. Journal of Social Psychology, 1941, 13, 123-140. MULL, H. K. The effect of repetition upon enjoyment of modern music. Journal oj Psychology, 1957, 43, 155-162. MUNSINGER, H. L. Meaningful symbols as reinforcing stimuli. Journal oj Abnormal and Social Psychology, 1964, 68, 665-668. National Opinion Research Center. Jobs and occupations: A popular evaluation. In R. Bendix & S. B. Lipsit (Eds.), Class, status, and power: A research in social stratification. Glencoe, 111.: Free Press, 1953. NEWCOMB, T. M. Stabilities underlying changes in interpersonal attraction. Journal of Abnormal and Social Psychology, 1963, 66, 376-386. O'CoNNELL, D. C. Concept learning and verbal control under partial reinforcement and subsequent reversal and non-reversal shifts. Journal of Experimental Psychology, 1965, 69, 144-151. OSGOOD, C. E., Suci, G. J., & TANNENBATJM, P. H. The measurement of meaning. Urbana: University of Illinois Press, 1957. PALERMO, D. S., & JENKINS, J. J. Word association normsgrade school through college. Minneapolis: University of Minnesota Press, 1964.

THE ATTITUDINAL EFFECTS OF MERE EXPOSURE


UNDERWOOD, B. J. Verbal learning and the educative process. Harvard Educational Review, 1959, 29, 107-117. VAN DER BEKE, G. E. French word book. New York: Macmillan, 1929. VAN DONGEN, G. A. Ameliorative! in English. Rotterdam: De Vries, 1933. VERVEER, E. M., BARRY, H., JR., & BOUSFIELD, W. A. Change in affectivity with repetition. American Journal of Psychology, 1933, 45, 130-134. WASHBURN, M. F., CHILD, M. S., & ABEL, T. M. The effects of immediate repetition on the pleasantness or unpleasantness of music. In M. Schoen (Ed.), The effects of music. New York: Harcourt, Brace, 1927.

27

WIEBE, G. The effects of radio plugging on students' opinions of popular songs. Journal of Applied Psychology, 1940, 24, 721-727. WILNER, D. M., WALKLEY, R. P., & COOK, S. W. Residential proximity and intergroup relations in public housing projects. Journal of Social Issues, 8, 1, 19S2. WLLSON, L. R., & BECKNELL, J. C. The relation between the association value, pronounciability, and affectively of nonsense syllables. Journal of Psychology, 1961, 52, 47-49. YELEN, D. R., & SCHULZ, R. W. Verbal satiation? Journal of Verbal Learning and Verbal Behavior, 1963, 1, 372-377. (Received April 11, 1967)

You might also like