Electronic Journal of Human Sexuality, Volume 13, November 11, 2010


Priming Effects Upon Memory For Double Entendre Words

James H. Geer1, Janice A. Gordon2, Alyssa Sharkey1, and Laura Thomas1

1 Department of Psychology, Franklin and Marshall College and 2 Selective Mutism Research Institute


Correspondence concerning this article should be addressed to James H Geer, F & M, Department of Psychology. Box 3003, Lancaster, PA, 17504-3003. Phone 717-291-4373. Fax: 717-291-4387.



Sixty-four undergraduate students were participants in a study that preceded double entendre (DE) words with either sexual or nonsexual primes.  Enhanced memory for emotional words has been well established (Kensinger & Corkin, 2003).  It was hypothesized that memory would be greater for DE words with sexual primes than nonsexual primes.  It was expected that the primes would provide a setting that would evoke either the sexual or nonsexual meaning of the DE word.  In support of the experimental hypothesis, recall memory for DE words with sexual primes was greater than for DE words in the nonsexual context.  The analysis of intrusions found that sexual intrusions were uncommon.  That finding has implications for the study of false memory in the domain of sexuality.  No significant gender effects  were found.  The present research provides a further empirical link between sex research and cognitive psychology. 


This paper reports a study that examines the effect of priming upon memory for double entendre (DE) words. There is significant literature demonstrating that emotional stimuli are better remembered than are nonemotional stimuli (Kensinger & Corkin, 2003). It has been argued that this effect serves an adaptive function. That is, the ability to more clearly remember emotional stimuli helps to prepare the individual for similar and likely important future experiences (Nielson & Powless, 2007). As an alternative explanation Talmi & Moscovitch (2004) suggest that organizational processes such as found in categories of words (for example, emotional) facilitate memory enhancement.

The finding that sexually explicit words are experienced as emotional may explain why they are remembered better than neutral words (Bush & Geer, 2001; Hadley & MacKay, 2006). Similar findings have been reported at a less detailed level in the study of memory for sexually explicit versus nonexplicit stories. In that research, explicit details are better remembered than nonexplicit details (McCall, Rellini, Seal, & Meston, 2007). Similarly, the concept of source memory in relation to emotional and nonemotional words has been studied. Doerksen and Shimamura (2001) in their study of source memory found that participants more accurately remembered surrounding contexts of emotional words than nonemotional words. Kensinger and Corkin (2003) found similar results. In one of their experiments they incorporated the use of arousal words instead of just emotional words. They defined arousal words as either sexual or swear words that were ranked higher on arousal than emotional words. They found that arousing words had greater recognition, recall, and source memory than emotional or neutral words.

Gender differences in lexical decisions concerning sexual words have been reported. In Geer & Bellard (1996) it was found that women’s identification of letter strings as words (lexical decisions) were delayed when presented with sexual words as compared with romantic words. This finding reveals an effect of meaning category on a basic level of information processing. In a similar finding McCall et al. (2007) reported that men recalled more erotic or explicit details of a story while women were more likely to remember love and emotional bonding details. Because of these findings of gender differences in the processing of sexual material, in the present research gender differences are examined.

In English many words have more than one meaning. Of interest to this work is the well-known fact that there are English words that have both sexual and nonsexual meanings. These words, typically referred to as DE, provide us with the possibility that the same word may vary in its memorability depending upon which of the meanings is activated by primes. By using the same word following different primes, it is possible to control for variables related to the target word’s characteristics, such as familiarity, length, pleasantness, emotionality and acceptability. By altering only the types of words that precede the DE words, two different contexts can be created while avoiding confounds that are associated with using different target words when studying two or more contexts. It was predicted that DE words, when primed by sexually explicit associates would be remembered more accurately than the same DE words that were primed by associates of the nonsexual meaning of the DE words.



Participants were 31 male and 33 female undergraduates who received course credit for participation in research. Participants were randomly assigned to the conditions of either the sexual priming list or the nonsexual priming list. Participants who were scheduled to receive sexually explicit words were so informed and told that declining to participate would not have any negative consequences. All participants agreed to participate upon being informed of the presence of explicit sexual words in the experiment. Prior to participation all participants signed an informed consent form that detailed the study. The fact that participants who were exposed to explicit sexual words were warned that such words were present presents a potential additional priming condition. It was our judgment that the likelihood of such a warning being an effective prime was limited.


The two word lists of 52 words each were constructed and were tested for differences in word length and frequency. Using t Tests to compare list types, it was found that the lists did not reliably differ (p > .05) on those variables. This should not be surprising as only 10 of the 52 words differed between the lists. Both lists began and ended with two neutral words in the attempt to reduce primacy and recency effects. Each word list contained the same five sets of related neutral words (ex. song, melody, music) as well as the five sets of words containing DE words and their primes. In addition to these 10 sets of words there were 22 neutral filler words that were dispersed evenly between the word sets.

As noted, identical neutral word sets and filler neutral words appeared in both lists. The two lists differ only in the words that preceded the DE words. In the sexual context list two sexual associates preceded each DE word. In the nonsexual list two nonsexual associates preceded each DE word. The word lists are found in the Appendix. We constructed three different random orders of each list to be sure that any identified effects were not due to the order of presentation. Participants were randomly assigned to one of the three variants of their list type.


After being briefed on the details of their participation and signing an informed consent form, participants sat in front of a computer monitor. Instructions included describing that participants would be asked to recall the word list. The experimenter also explained that words would be presented on the computer screen. Each word was presented for 1500 ms with 500 ms between each word. All participants were told that some words would appear in clusters. As noted, participants who were assigned to see the list with sexual words were advised that such words would be present. After the experimenter started the program, she left the room and shut the door. Following the last word the message “Please knock on the door for the experimenter.” appeared. When summoned, the experimenter reentered the room. A five-minute letter search task that served to give the participants an activity that interfered with rehearsing the word list was initiated. The task was to look at random letters that were typed onto pages and to cross out all vowels. The experimenter left the room while the participant worked on the letter search. The experimenter signaled when five minutes had passed.

Participants were then given a blank sheet of paper and asked to write down as many words as they could remember from the presented list. There was no time limit. After the participants completed the recall task they were thanked and asked to leave their email address if they wanted the results of the study. Also, after the participant departed, their assigned experimental number, gender, and type of list (sexual or nonsexual) was recorded on their recall sheet. Note, to insure confidentiality, the subjects’ names were not placed on the data sheets.


The design of the study is straightforward. There are two principal independent variables: list type (sexual primes vs. nonsexual primes) and participant gender. The principal dependent variables examined were the number of words correctly recalled for each of the various word categories. In addition the number and type of intrusions or false memories that occurred was recorded and analyzed.

The principal analysis is of the recall of DE words. These data were examined by a 2 x 2 univariate ANOVA. The results of that analysis fulfilled our expectation that DE words in the sexual list would be recalled at a higher rate. The mean correctly recall DE words in the sexual list was M = 2.52 which was significantly greater, F(1,60) = 6.864, p = .011, Partial Eta 2 = .103, than the mean M = 1.81 for recalled DE words in the nonsexual list. Out of the 52 presented words, only five were double entendre. Only 24% of the five were recalled when preceded by neutral primes while 50% were recalled when preceded by sexual primes. This finding supports our prediction that recall memory with its preceding encoding for DE words differ depending upon priming context within which they are placed.

Several analyses were conducted to examine other word types. There were two word types used as primes that preceded each DE word. In the sexual list those primes were sexual associates of the DE words. In the nonsexual list, the primes were nonsexual associates of the DE words. We predicted that the sexual primes in the sexual list would be recalled at a higher frequency than the nonsexual primes in the sexual list. The univariate ANOVA of primes found a significant effect of list type, F(1,60) = 99.461, p < .001, Partial Eta 2 = .642. The mean number of primes that preceded the DE words that were correctly recalled for the sexual list was M = 6.21 and for the nonsexual list the mean was M =2.19. Gender and the interaction terms were not statistical significant. As these data show, sexual words in the prime position were recalled at a much higher frequency than nonsexual words. This, of course, is consistent with the previously reviewed literature showing enhanced memory for emotional words.

Two word categories, filler and buffer, were used to enlarge the list as pilot work demonstrated that a shorter (15 item list) suffered from a ceiling effect. However, we did not expect these neutral words to be influenced by either the list type or participant gender. There were the 18 unrelated nonemotional filler words that were placed in five sets of three that occurred before each of the five DE sets and one set of three that occurred after the last DE set. There were two neutral buffer words at the beginning and end of the list to reduce primacy and latency effects. These buffer words were not included in our analysis of nonemotional words. In a 2 x 2 univariate ANOVA of the filler words there were no gender or list effects that were statistically significant. The interaction term also failed to demonstrate significance. These results indicate that the presence of sexual words or the presence of warnings that sexual words would be presented did not reliably influence the recall of unrelated neutral words present in the list.

Our final analysis is of intrusions. Intrusions were words that participants listed during the recall task that had not appeared in their list. We categorized intrusions into the categories sexual and nonsexual. To analyze intrusions we intended to conduct a MANOVA with nonsexual intrusions being one dependent variable and sexual intrusions being the other. However, inspection of the data revealed that there were no sexual intrusions for women and only four for men. The lack of variance for sexual intrusions makes it unwise that a MANOVA be conducted.

While a MANOVA seems inappropriate, the intrusion data clearly pointed out that there were more nonsexual intrusions than sexual intrusions. The mean for sexual intrusions was M = .06 and for nonsexual intrusions the mean was M = 2.11. This difference was significant, t(63) = 7.544, p < 001. A possible explanation of this effect is that participants were reluctant to report sexual words having been present when there is some question of the report’s accuracy. It is also true that there were more nonsexual than sexual words in the lists. Perhaps that facilitated the incorrect recall of nonsexual words.


Our study demonstrated that DE words are recalled at different levels depending upon which of the word’s meanings has been activated. That is, recall memory for DE words preceded by sexual primes is greater than recall of the same words preceded by nonsexual primes. The finding that activating the sexual meaning of DE words by the presence of sexual primes increases the memorability of DE words illustrates how sexual meaning influences fundamental cognitive processes. The findings show the relevance of experimental cognitive research to the domain of sexuality. A perusal of the literature on sex research reveals a scarcity of references to the burgeoning cognitive literature. Closer ties between basic experimental research and sexuality research would be a value to both domains.

The importance of studies using domains typically not involved in cognitive psychology is noted by Sternberg (2006). In his text on cognitive psychology Sternberg notes that one of the “Underlying Themes in the Study of Cognitive Psychology" is “Domain generality versus domain specificity" (p. 20). That is, “Are the processes we observe limited to a single domain, or are they general across a variety of domains" ( p. 20). The current research fits well under that rubric by demonstrating how priming and memory for sexual words follows patterns familiar to the scholar of cognitive psychology working with nonsexual stimuli.

Analyses revealed that participants recalled significantly more sexual priming words than nonsexual priming words. These data are consistent with previous work showing that sexual words are recalled at a higher frequency than nonsexual control words. Those data are also consistent with the general finding from cognitive psychology that emotional words yield enhanced memory (Kensinger & Corkin, 2003.) Apparently sexual words are more salient than nonemotional neutral stimuli and thus are recalled at a greater level. Our study does not, however, test alternative explanations for the saliency of sexual words. These finding, once again, reveal that findings in the domain of sexuality are consistent with findings from basic cognitive psychology. This argues further for the incorporation of concepts and paradigms from basic cognitive research into models and paradigms when studying sexuality.

There were very few sexual intrusions, only a total of four. On the other hand, there were over 120 nonsexual intrusions. Not counting the DE words as sexual, there were approximately 5 times the numbers of nonsexual to sexual words in each of the two lists. However, there were more than 40 times more nonsexual than sexual intrusions. These data may reflect the saliency of sexual words. That is, sexual words, being more salient are more readily recalled and, as a result, incorrect alternatives may be more readily identified. Or perhaps the difference may simply due to the fact that participants were presented with fewer sexual than nonsexual words. The data do suggest that sex researchers need to be careful when examining false memory in sex research. Our finding concerning the scarcity of sexual intrusions suggest the possibility that research participants may be more reluctant to report sexual memories for which they are uncertain than uncertain memories from other domains. Or perhaps false memories are simply less common in the domain of sexuality. The answer awaits further study.

In summary, our findings reinforce the relevance of cognitive theory and findings to the study of sexuality. Our data show that the basic phenomenon of priming in cognitive psychology is applicable to sexuality. Priming was shown to affect the basic phenomenon of memory in that DE words were remembered better when their sexual meaning was primed. The finding of fewer sexual than nonsexual false memories suggests that researchers studying false memory need to be cautious when working with sexual stimuli. Our false memory findings suggest that it is possible that the criteria employed by participants in reporting the memory of a sexual stimulus is particularly stringent.


Bush, S.I., & Geer, J.H. (2001). Implicit and explicit memory of neutral, negative emotional, and sexual information. Archives of Sexual Behavior, 30, 615-630.

Doerksen, S. & Shimamura, A.P. (2001). Source memory enhancement for emotional words. Emotion, 1, 5-11.

Geer, J.H., & Bellard, H.S. (1996). Sexual content induced delays in unprimed lexical decisions: Gender and context effects. Archives of Sexual Behavior, 25, 379-396.

Hadley, C.B., & MacKay, D.G. (2006). Does emotion help or hinder immediate memory? Arousal versus priority-binding mechanisms. Journal of Experimental Psychology: Learning, Memory, and Cognition, 32, 79-88.

Kensinger, E.A. & Corkin, S. (2003). Memory enhancement for emotional words: Are emotional words more vividly remembered than neutral words. Memory and Cognition, 31, 1169-1180.

McCall, K.M., Rellini, A.H., Seal, B.N., & Meston, C.M. (2007). Sex differences in memory for sexually-relevant information. Archives of Sexual Behavior, 36, 508-517.

Nielson, K.A., & Powless, M. (2007). Positive and negative sources of emotional arousal enhance long-term word-list retention when induced as long as 30 min after learning. Neurobiology of Learning and Memory. 88, 40-47.

Sternberg, R. J. (2006). Cognitive Psychology, Fourth Edition. Belmont, CA. Thompson Higher Education.

Talmi, D. & Moscovitch (2004). Can semantic relatedness explain the enhancement of memory for emotional words? Memory & Cognition. 32 (3), 742-751.


Words Used In The Study

Sets of DE words with their preceding sexual associates

Testicles Genitals Balls     Intercourse Fuck Screw     Penis Dick Prick

Vagina Clitoris Pussy    Condom Trojan Rubber

Sets of DE words with preceding nonsexual associates

Bats Base Balls    Hammer Wrench Screw    Stab Thorn Prick

Kitty Puppy Pussy    Plastic Vinyl Rubber

Sets of related neutral words

Pistol Rifle Gun    Song Melody Music    Skeleton Bone Rib

Owl Robin Hawk    Pineapple Banana Orange

Neutral filler words

Broth  Coleslaw  Levees  Igloo  Airstrip  Skyscraper  Dune  Monument  Mosques  Coals  Dusk  Rainbow  Fried  Comedy  Mule  Podium  Atom  Clouds

Words used to reduce primacy

Swift  Dock

Words used to reduce recency

Bread  Magic

Return to Front Page