Expect the Unexpected? Testing for Minimally Counterintuitive (MCI) Bias in the Transmission of Contemporary Legends

Abstract

In this study, we use computational methods for analyzing cultural transmission to examine the role of cognitive selection pressures on the evolution of narratives, the first use of computational phylogenetic analysis in the study of contemporary legends. It has been suggested that a number of biases in transmission may alter the content and structure of narrative so as to maximize how transmittable it is. One bias that has attracted much attention is Boyer’s minimally counterintuitive (MCI) bias, which suggests that a cognitively optimal number of counterintuitive concepts increase the salience and, therefore, the transmission of a narrative. Previous research has used traditional folklore and religious texts to examine this bias and a cognitively optimum number of 1–2 or 2–3 counterintuitive concepts has been suggested. The present research uses the legend of “Bloody Mary,” a contemporary (or urban) legend with MCI elements in a computational phylogenetic analysis to examine the influence of MCI on cultural transmission and evolution. Counterintuitive and intuitive concepts were found to be equally stable in transmission, suggesting that MCI bias may function on the narrative as a whole, rather than individual concepts within it.

Keywords

computational phylogenetic analysis cultural transmission evolution of narratives urban legends

Introduction

In recent years, the study of cultural transmission—that is, the preservation and modification of socially learned information––has been greatly enhanced by the application of computational quantitative methods, such as mathematical modeling, simulations, and phylogenetic software. The introduction of these techniques has allowed the investigation into the dynamics of cultural evolution to develop from a largely theoretical or philosophical exercise to a fully fledged empirical research field (see Mesoudi, 2011). In this article, we adopt computational phylogenetic methods to investigate the potential role played by psychological biases on evolution of narratives. It has been proposed that when stories are repeatedly retold from one person to another they are subjected to cognitive selective pressures which may alter the content and structure so as to make them maximally transmittable (Barrett & Nyhof, 2001). Rather than a process of high fidelity replication, human cultural transmission displays biases and distortions which reflect these cognitive selection pressures. Biases and distortions in recall were first identified by Bartlett (1932) and since then studies have identified a number of potential biases in transmission. One such bias was suggested by Boyer’s (1994) theory of the transmission of counterintuitive ideas, also termed minimally counterintuitive (MCI) bias.

MCI Bias

Boyer (1994) argues that for any category, people hold a host of intuitive assumptions about its member’s properties. These intuitive assumptions are generally described as “folk biology,” “folk physics,” and “folk psychology.” These intuitive assumptions do not involve conscious inspection and are assumed to be cross-cultural (Barrett, 2008). Concepts which violate these category-level expectations are considered to be counterintuitive and feature inherent transmission advantages that can increase the salience of a narrative (Boyer, 1994). Boyer, however, also argues that this effect is not linear and that a balance must be met between satisfying intuitive expectations and violating enough of them to be cognitively optimal. If a narrative features too many counterintuitive concepts it can render it incomprehensible and unmemorable (Boyer & Ramble, 2001). Stories which are considered MCI will balance a minority of counterintuitive concepts with a majority of everyday, intuitive events. Several studies have supported the idea that MCI narratives are cognitively optimal, granting them an advantage in recall and transmission (e.g. Barrett, Burdett, & Porter, 2009; Barrett & Nyhof, 2001; Boyer & Ramble, 2001; Norenzayan, Atran, Faulkner, & Schaller, 2006; Upal, 2011; Upal, Gonce, Tweney, & Slone, 2007). This content bias has been explained in evolutionary terms by Upal et al. (2007), who argues that memory should evolve to preferentially remember and recall events or objects which violate expectations about the future but can be justified once they have been observed. While Barrett (2000) suggests that MCI concepts hold a transmission advantage because they can be understood and represented without the allocation of too many cognitive resources but are also challenging enough to require increased attention in order to be assimilated into conceptual schemes

Barrett and Nyhof (2001) used a transmission chain methodology to examine whether MCI narratives have a cognitive advantage. Using both traditional Native American folktales (as Bartlett, 1932, had used) and a science-fiction style story, they found that, after three generations of transmission, counterintuitive, and intuitive but bizarre items were recalled in greater proportions than intuitive items. This recall of counterintuitive items was found both immediately and after a 3-month delay. They also replicated these findings using oral transmission in a group serving as a mock community. The fact that recall of counterintuitive items was upheld after a three month gap is important, as an idea that is immediately memorable but becomes significantly less memorable with time could not be as culturally successful as those ideas that retain memorability.

Norenzayan and colleagues (2006) conducted two studies using different experimental methodologies other than transmission chains. The first study used lists featuring intuitive and counterintuitive ideas and, while the intuitive ideas were better recalled in the short-term (similarly to the results of Bartlett, 1932) MCI lists had a recall advantage over time. For their second study, they used the folktales collected by the Brothers Grimm and had the rated for MCI concepts, memorability, and other psychological variables. They found that MCI tales were more likely to be culturally successful (determined by number of hits using a Google web search) and were perceived to be more memorable and more psychologically appealing in terms of understandability and ease of transmission. They suggest that the cognitively optimum level is around two to three counterintuitive objects in a narrative. Norenzayan et al. (2006) concluded that MCI narratives are culturally successful due to a stronger cognitive advantage in recall compared to other narrative templates.

Barrett (2008) acknowledges some of the issues relating to the empirical study of MCI bias. Identifying a difficulty in ascertaining what are and are not counterintuitive concepts in a technical sense and an inability in quantifying just how counterintuitive a concept is. In an attempt to clarify MCI theory and aid the future empirical study of MCI concepts, Barrett (2008) presented a formal system of coding and quantifying the “counterintuitiveness” of a concept termed the counterintuitiveness coding and quantifying scheme (CI-Scheme). The CI-Scheme uses six steps which include identifying basic level membership and ontological category/categories of a concept and coding for breaches and transfers of expectation sets. Using this scheme, a concept can be given a “counterintuitiveness score.” Barret (2008) suggests that MCI concepts are those which score one for counterintuitiveness and that the cognitive optimum would be a score of one or two, concepts which score three or higher would too counterintuitive to have a transmission advantage and would be discarded for simpler concepts. Barrett et al. (2009) uses the CI-Scheme in a study which analyzed a cross-cultural selection of 73 folktales. Independent coders applied the CI-Scheme to these folktales and it was found that the number of counterintuitive objects ranged from 0 to 11, however, the majority (79.2%) only contained one or two objects. Most of the counterintuitive objects (99%) had a score of one or two, none had a score higher than three. Barrett (2009) uses these findings to argue that the cognitive optimum for the number of counterintuitive concepts in folktales is one or two and that these objects will only violate one or two ontological categories.

MCI in Contemporary Legends

Previous research into MCI bias has used traditional folklore or religious texts as a means to study its effect on transmission and memorability, however, examples of MCI can also be found in an area of folklore which is less “canonized”; the contemporary legend. Contemporary legends, also referred to as “modern legends,” “urban belief tales,” and “urban legends” are generally defined as apocryphal stories which are told as true (Brunvand, 2000; Heath, Bell, & Sternberg, 2001; Tangherlini, 1990), involve a contemporary setting (Brunvand, 2000), and feature a single event, usually an individual experience, as the core of the narrative (Tangherlini, 1990). Successful legends often share a number of features, such as a suspenseful or humorous narrative (Brunvand, 2000) that contains surprising information or a twist ending (Fox Tree & Weldon, 2007), a warning or moral message that is either explicit or implied, and they are often attributed to a “friend of a friend” (Brunvand, 2000). While they have been traditionally transmitted orally, contemporary legends are now spread through a combination of oral transmission, electronic communication, and publication in mass media (Brunvand, 2000). The analysis of contemporary legends can offer a unique means of studying the concerns of modern populations (Brunvand, 2000) and therefore can be a fruitful area to study content biases such as MCI bias.

One example of a contemporary legend which includes MCI concepts is the legend of “Bloody Mary.” A wide number of variations of this legend exist but generally the story involves a young girl (the character is female in almost all examples) repeating a certain phrase a specific number of times before a mirror. As a result of the ritual, a frightening, supernatural figure appears and often inflicts violence upon the summoner. Below is a typical example:

A long time ago there was a woman named Mary. She was very beautiful and very vain. She had a horrible accident though and her face became so horribly scratched that she bled to death. Her spirit didn't die, she still comes; if you stand in front of a mirror in the dark and say her name three times her face appears. If you do'nt [sic] turn on the light and run away as fast as you can she'll scratch your face off, or some say, drag you into the mirror with her!

-Collected from http://www.oocities.org/tragicpixie/FLbloodymary.html.

The name and nature of the supernatural figure can vary but most frequently it is a ghost named Mary, with Bloody Mary being the most common variation. Other variations include Mary Worth, Mary Whales, Black Agnes, the White Lady, the Bell Witch (a regional variation specific to Tennessee, USA), and Svarte Madame. Most commonly the summoning phrase is simply the figure’s name, and typically it is repeated 3 times. The figure is sometimes given a back-story, such as being the ghost of a woman executed for witchcraft or a beautiful woman disfigured. If present, the back-story often links to the summoning phrase, for example, the summoning phrase “Bloody Mary, I killed your baby” is often linked to the back-story of Bloody Mary being the ghost of a mother whose children were murdered. The result of the summoning also varies; sometimes the summoner is killed or injured by the supernatural figure, while in other examples a means of escape or protection is present, such as turning on a light or a protective ring of salt.

The Bloody Mary legend are popular and widespread, with similar variations being found in Britain, the United States, Canada, Sweden, Holland, France, Australia, Japan, and Thailand (Armitage, 2006). It has also found its way into popular culture, being the focus of horror films such as Urban Legends: Bloody Mary (Merrell, Phillips, Messer, & Lambert, 2005), Bloody Mary (Miller, Tyler, & Valentine, 2006), Dead Mary (Cameron et al., 2007), The Legend of Bloody Mary (Ahrens, Locke, Sawyer, & Stecenko, 2008), and was an influence on the film Candyman(Barker, Golin, & Rose, 1992).

Previous research has suggested that the Bloody Mary legend evolved from much older mirror divination rituals (Armitage, 2006; Klintberg, 1988). In his notes on his 1786 poem Halloween, Burns (2009) describes a mirror ritual that a woman can use to see the face of her future husband. This was a popular piece of European folklore from the late 18th to 20th century (Hutton, 2001), with Halloween greetings cards being produced with pictures of the ritual. The results of the ritual could, however, be horrific. It was said that if the woman was destined to die unmarried she would see the face of the Grim Reaper reflected in the mirror (Ellis, 2004). The 19th-century ballad Svetlanaby Vasili Zhukovskii, an example of “folklore” romanticism (Ryan & Wigzell, 1992) also features the accidental summoning of a horrific supernatural figure; Svetlana is persuaded by her female companions to enact a mirror ritual to see her lover, however, a demon appears in his place. It was also said during the 19th century that if you stare into a mirror for too long the devil would appear (Jones & Kropf, 1883). These examples of traditional folklore are likely to be the ancestors of the modern Bloody Mary legend. It was not until the 1970s, however, that examples of Bloody Mary legends were first published in print. The first examples of “Mary Worth” legends were published in an anthology of American children’s folklore (Knapp & Knapp, 1976) and the first academic writing on the subject was by Longlois (1978). As with any oral tradition, Bloody Mary legends could have existed for many years before they were first published, however, it is argued that the Bloody Mary legend originated in the USA (Klintberg, 1988).

Computational Phylogenetic Analysis

The current research uses computational phylogenetic analysis techniques to analyze variations of the Bloody Mary legend. Phylogenetic analysis employs an explicit evolutionary model whereby new entities arise by descent with modification from existing ones. In this context, phylogenetic analysis provides a useful means for estimating relationships of common ancestry among story variants, and examining which traits tend to be preserved and which ones modified. This allows us to explore the influence of MCI elements on the cultural evolution and transmission of a narrative and will particularly focus on whether the inclusion of an MCI element makes the narrative as a whole more stable, or whether the individual elements vary in stability, with the most stable being those which are MCI. Based on the work of Barrett and Nyhof (2001) and Norenzayan et al. (2006), one would expect the MCI concepts to display greater stability in transmission than the intuitive concepts of the narrative.

The software we used for this analysis was Phylogenetic Analysis Using Parsimony (PAUP) (Swofford, 2003), a widely used software package for the inference of evolutionary trees that supports a wide range of approaches to phylogenetic analysis and features relatively friendly input for data and output of results. It is proprietary software, but an open source package PHYLIP (PHYLogeny Inference Package, Felsenstein, 1989) is also in common use.

Method and Materials

Legend variations were gathered from internet websites (see Appendix A for a full list of sources) using the search key words bloody mary. For the purposes of this study, any contemporary legend which features the purposeful summoning of a supernatural, malevolent figure through a ritual involving a mirror is considered a Bloody Mary legend variation and was included in the dataset. Fragmentary legends were excluded. 45 Bloody Mary legend variations were collected and a total of 36 characters were derived from the sample. These characters included features of the supernatural figure such as appearance, elements of the ritual such as summoning phrase and number of repetitions and other key elements of the narrative (see Appendix B for full list of characters). Character states were coded as present or absent (see Table 1).

Table 1.

Representative Characters and Coding (see Appendix B for full list).

Character	Trait	Coded As
1	Mary	0 = Absent;1 = Bloody Mary;2 = Mary Worth
2	Satan or Devil	0 = Absent;1 = Present
3	Bathroom	0 = Absent;1 = Present
4	Name summons	0 = Absent;1 = Present
5	Belief summons	0 = Absent;1 = ‘I believe in’;2 = I don’t believe in’
6	“Come to/get me” summons	0 = Absent;1 = Present
7	Baby summons	0 = Absent;1 = Present
8	Number of phrase repetitions	0 = Absent;1 = Three;2 = Seven;3 = Thirteen;4 = Twenty;5 = Unspecified;6 = Five;7 = Six;8 = Fifty; 9 = Ten
9	Beautiful woman disfigured	0 = Absent;1 = Present
10	Mother of murdered children	0 = Absent;1 = Present

In order to examine the hypothesis that MCI concepts within a narrative feature a mnemonic advantage, characters were coded as either counterintuitive or intuitive. Characters were coded as counterintuitive if they breach the folk expectations of the ontological categories suggested by Boyer (1994). Characters were only coded as counterintuitive if the “counterintuitiveness” was a property of that individual character in itself; characters which brought about counterintuitive events in the context of the narrative but did not breach folk expectations themselves, such as the ritual summoning phrase, were therefore considered intuitive (see Barrett, 2008). For example, the ritual taking place in a bathroom (Character 3) would not be coded as counterintuitive, as its presence in the narrative does not breach the folk expectations of any ontological categories, however, the supernatural figure appearing through a mirror or out of a bath or sink (Character 16) would be coded as counterintuitive, as it breaches the expectations of “folk physics.”

The character data were analyzed using cladistic phylogenetic reconstruction. Cladistic analysis reconstructs relationships by minimizing the number of evolutionary changes required to explain the distribution of character states among a group of taxa. This involves distinguishing characters that are evolutionarily novel (also termed apomorphic or derived), from those that were present in the last common ancestor of all the taxa under study, which are labeled ancestral or plesiomorphic. The presence of a derived trait in two or more taxa provides evidence that they are descended from a common ancestor of more recent origin than the ancestors they share with the other taxa under analysis. There are several methods to identify which traits are derived and which are ancestral, the most popular of which is outgroup analysis. An outgroup is defined as a taxon that shares a common ancestor with the taxa under analysis (the ingroup), but is of more distant origin than the ancestor the analyzed taxa share with each other. Since the outgroup does not share an exclusive common ancestor with any individual member of the ingroup, it follows that when a character occurs in two states among the study group, but only one of the states is found in the outgroup taxon, the former is considered the derived state and the latter the ancestral state. For the purposes of these analyses, the variant which most resembles traditional mirror ritual folklore (T42) was used as an outgroup.

Although the principles of cladistics are relatively simple, actually identifying the shortest, or most parsimonious tree (MPT) is usually impossible without the aid of computers as the number of possible trees for any given dataset increases exponentially with the number of taxa. For three taxa (A, B, C), we only need to compare the number of steps for three possible trees (AB-C; AC-B; CB-A). But for four taxa we must consider 15 trees, for seven taxa 10,000 trees, and for 10 taxa we would have to evaluate more than 34 million trees. Therefore, most cladistic analyses require specialized software programs that use algorithms to search for the MPTs. In this study, we used the heuristic branch-swapping search function in the program PAUP 4.0 (Swofford, 2003).

Once phylogenetic relationships among the variants were estimated, the fidelity of transmission from ancestral to descendent taxa was estimated using a statistic known as the Retention Index (RI). The RI is a measure of the number of homoplastic changes a cladogram requires that are independent of its length (Farris, 1989). The RI of a single character is calculated by subtracting the number of character state changes required by the focal cladogram (s) from the maximum possible amount of change required by a cladogram in which all the taxa are equally closely related (g). This figure is then divided by the result of subtracting the minimum amount of change required by any conceivable tree (m) from g. The RI of two or more characters is computed as (G − S)/(G − M), where G, S, and M are the sums of the g, s, and m values for the individual characters. A maximum RI of 1 indicates perfect fidelity of character transmission, while the lower the fidelity the closer the RI reaches 0. The RI was used to estimate both the overall fidelity of transmission of characters, and to compare counterintuitive and intuitive characters.

Results

The number of counterintuitive characters found in the legend variations ranged from 1 to 4 (M = 2.356, SD = .883) and the number of intuitive characters found in the legend variations ranged from 1 to 10 (M = 5.756, SD = 2.176). 91.12% of legend variations featured between one and three counterintuitive concepts.

The cladistic analysis in PAUP returned 147 MPTs (i.e., there were 47 trees that required an equal number of evolutionary changes). The character RIs of two consensus trees were used for analysis. One tree represented relationships that occur in the majority of the MPTs, referred to as the Majority Rules Tree (Figure 1), and another tree represented relationships that occur in all of the MPTs, referred to as the Strict Tree (Figure 2).

Figure 1.

Strict consensus of the most parsimonious trees returned by the cladistic analysis.

Figure 2.

Majority rules consensus of the most parsimonious trees returned by the cladistic analysis.

The RI of all the characters on the two consensus trees was .53. As the data were not normally distributed, a Mann–Whitney U test was conducted to compare the individual RIs of counterintuitive and intuitive characters. No significant difference was found in either the Majority Tree (U = 85, z = −1.624, p > .05) or the Strict Tree (U = 80, z = −1.817, p > .05). The frequency of appearance in the variations of counterintuitive and intuitive characters was also compared using a Mann-Whitney U test but no significant difference in frequency of appearance was found (U = 105, z = −0.894, p > 0.05).

Discussion and Conclusions

The RIs of all the characters on the two consensus trees are considered low by the standards of other real and simulated cultural data sets (Collard, Shennan, & Tehrani, 2006; Nunn, Arnold, Matthews, & Borgerhoff Mulder, 2010). This indicates that the Bloody Mary story is transmitted with generally poor fidelity from ancestral to descendent version. One explanation for this finding concerns the nature of the legend itself and how it has traditionally been transmitted. For at least 30 years, Bloody Mary legend variations have been more often transmitted orally in a group setting rather than between individuals (Dundes, 1998) suggesting that the most likely form of transmission is nonlinear. Rather than one individual passing on their version of the legend to another individual, a group of people may bring together different versions and elements of all of them may be combined and transmitted. At present there has been little research into the effects of group transmission on the cultural evolution of a narrative, the most commonly used experimental paradigm has been the linear transmission chain, where an individual transmits a single narrative to another individual (Eriksson & Coultas, 2012). Barrett and Nyhof (2001) and Marques, Páez, Valencia, and Vincze (2006) both used experimental designs in which multiple cultural parents were used but no analysis was conducted to compare this with a traditional linear design. Eriksson and Coultas (2012) did examine the influence of a two cultural parent design, where each participant receives transmission from two separate sources, and found that recall was greater in the two cultural parent chains than single or double transmission chains. Items which were present in both versions of the story were most stable in transmission and items which were unique to the first story heard had a recall advantage over items which were unique to the second story. How group transmission where versions are not presented simply one after the other and where items may directly compete with each other, as they may between Bloody Mary legend variations, effects the transmission and evolution of a narrative has not yet been explored but it could explain for the weak phylogenetic signal found in these trees.

As the Bloody Mary variations were collected from the internet rather than oral sources, it is possible that this effected their transmission. Unlike oral transmission the electronic transmission of a narrative allows people to copy exactly what has been sent to them, without the limitations of recall. As a means to check if the Bloody Mary legend variations used in analysis appeared to be copied from one another, the variations were entered into Turnitina computer program designed to compare academic essays and check for plagiarism. None were found share more than 10% similarity except for two pairs which shared around 50%, however, these pairs appeared to just be close variations rather than the result of direct copying. As a result it was deemed that these variations were likely to have been largely transmitted orally rather than electronically.

Norenzayan et al. (2006) and Barrett (2008) both suggest a fixed cognitive optimum number of counterintuitive concepts to be found in a story. The range of counterintuitive characters present in the Bloody Mary legends was found to be one to four, with an average of 2.36. This average falls within the cognitive optimum of two to three counterintuitive concepts as suggested by Norenzayan et al. (2006) and comes close to the one to two cognitive optimum suggested by Barrett (2008), the majority of stories (91.12%) featured one to three counterintuitive concepts with very few featuring four, none featured more than this. The research conducted by Norenzayan et al. (2006) and Barrett (2008) used traditional folktales, such as those collected by the Grimm Brothers, to derive their cognitive optimum and suggest that it would be true of all stories. The results of the current research suggest this may be the case, as a similar number of counterintuitive concepts were found in variations of a contemporary legend. It also suggests that contemporary legends and traditional folktales may feature structural similarities in terms of MCI content.

The finding that there was no difference in the RIs or appearance frequency of counterintuitive and intuitive characters, however, suggests that the counterintuitive characters were as stable in transmission as the intuitive characters and, therefore, may not be inherently more memorable as individual concepts. Should counterintuitive concepts feature an inherent mnemonic advantage, one would expect them to be more stable in transmission and the RI of counterintuitive characters would be significantly higher than intuitive characters. As this was not the case it suggests that the transmission advantage of MCI bias works as a function of the narrative as a whole rather than the individual concepts within it. Although it could be argued that the weak phylogenetic signal in both sets of characters means that we should not read too much into these findings, it is worth recalling that a comparison of the raw frequencies of counterintuitive and intuitive characters also found no significant difference. Again, this contradicts the prediction that counterintuitive traits are more likely to be copied than intuitive ones.

These findings are supported by Upal (2011) and Upal et al. (2007) who refers to the hypothesis that the presence of a specific number of MCI concepts increases the memorability of a narrative as the Fixed MCI-narrative template (FMNT) hypothesis and conducted a series of studies intended to critically examine this. Upal (2011) found that the inclusion of counterintuitive concepts only increases the memorability of a narrative if the concepts make the story more coherent. Upal (2011) suggests that the global cohesion of the narrative is a key mediating factor in its memorability and calls into question that idea that there is a specific cognitive optimum for all narratives. Upal also argues that memorabilty is not a property of a concept but it is a property of the concept, its context, and the background knowledge of the comprehendor, therefore the memorability of a concept will vary depending upon its context and this context should not be ignored when examining transmission (Upal, 2011). The importance of the context in which counterintuitive concepts are presented is also supported by Atran and Norenzayan (2004) who found that recall for counterintuitive concepts only increased when they were surrounded by intuitive concepts, suggesting that the counterintuitive concepts were not inherently more memorable, they only became more memorable when placed in an intuitive context. The interaction between counterintuitive and intuitive concepts within a narrative and their joint influence on its cultural transmission is an area that is yet to be fully explored.

In conclusion, this study demonstrates that contemporary legends provide rich data for studying processes cultural transmission. Previous research analyzing contemporary legends has primarily been focused on psychoanalytic interpretation of the legends (e.g., Dundes, 1998; Tucker, 2005) and while examples exist (see Heath et al., 2001), relatively little research has utilized these legends to investigate how narratives and other cultural forms might be influenced by evolved cognitive biases such as MCI bias. The phylogenetic analyses of Bloody Mary presented here represent significant steps in that direction and, we hope, further underline the important role that computer based techniques can play in the study of cultural transmission and evolution.

Footnotes

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was funded by the Health Research Council of New Zealand.

Appendix A

List of Legend Sources

www.allpoetry.com/contest/638007-blood_covered_mirror__

www.americanfolklore.net/folklore/2009/10/bloody_mary.html

www.halloween-website.com/bloodymarystories.htm

www.halloween-website.com/bloodymarystories2.htm

www.iusedtobelieve.com/make_believe/ghosts/bloody_mary/

www.oocities.org/tragicpixie/FLbloodymary.html

www.project2067.com/Urban%20Legends/Classic%20Horror/Bloody%20Mary.htm

www.retrocrush.com/100monsters/69.html

www.sfogszero.com/stories/bloody-mary

www.snopes.com/horrors/ghosts/bloodymary.asp

www.soccerpluscrap.8.forumer.com/viewtopic.php?t=196

www.unsolvedmysteries.com/usm22697.html

www.urbanlegends.about.com/od/horrors/a/bloody_mary.htm

www.urbanlegends.about.com/u/ua/horrors/your_favorite_scary_urban_legends.03.htm

www.urbanlegends.about.com/u/ua/horrors/your_favorite_scary_urban_legends.04.htm

www.web.archive.org/web/20020602081747/ http://www.mythologyweb.com/bloodymary.html

Appendix B

List of Characters and Coding

Character	Trait	Coded As
1	Mary	0 = Absent;1 = Bloody Mary;2 = Mary Worth
2*	Satan or Devil	0 = Absent;1 = Present
3	Bathroom	0 = Absent;1 = Present
4	Name summons	0 = Absent;1 = Present
5	Belief summons	0 = Absent;1 = ‘I believe in’;2 = I don’t believe in’
6	“Come to/get me” summons	0 = Absent;1 = Present
7	Baby summons	0 = Absent;1 = Present
8	Number of phrase repetitions	0 = Absent;1 = Three;2 = Seven;3 = Thirteen;4 = Twenty;5 = Unspecified;6 = Five;7 = Six;8 = Fifty; 9 = Ten
9	Beautiful woman disfigured	0 = Absent;1 = Present
10	Mother of murdered children	0 = Absent;1 = Present
11*	Appearance of supernatural figure	0 = Absent;1 = Facially disfigured;2 = Scissors in neck;3 = Headless;4 = Eyeless;5 = Covered in blood;6 = Old
12	Fingernail scratches injury	0 = Absent;1 = On Limbs;2 = On Body;3 = On Face
13	Face scratched off	0 = Absent;1 = Present
14	Clenched fist draws blood	0 = Absent;1 = Present
15	Spinning or turning around	0 = Absent;1 = Present
16*	Supernatural figure appears from	0 = Absent;1 = Mirror;2 = Bath/Sink/Toilet
17	Ritual takes place at midnight or specific time	0 = Absent;1 = Present
18	Silence from room	0 = Absent;1 = Present
19	Flushing toilet	0 = Absent;1 = Present
20	Summoner dies or killed	0 = Absent;1 = Present
21*	Room goes cold	0 = Absent;1 = Present
22*	Something appears on mirror	0 = Absent;1 = Blood;2 = Writing
23	Holding weapon	0 = Absent;1 = Knife;2 = Axe;3 = Hooked Hand
24*	Fingernails reach out of mirror	0 = Absent;1 = Present
25	Holding baby	0 = Absent;1 = Dead baby;2 = ‘Live’ baby;3 = Doll
26*	Actions predicted by colour	0 = Absent;1 = Present
27	Water thrown on mirror	0 = Absent;1 = Present
28	Test to see if figure attacks you	0 = Absent;1 = Given baby;2 = Asked question(s)
29	Candle(s) used in ritual	0 = Absent;1 = Present
30*	Means of protection or escape	0 = Absent;1 = Turning on light;2 = Ring of salt
31	Holding rose	0 = Absent;1 = Present
32	Candle is blown out	0 = Absent;1 = Present
33*	Blood flows from sink/bath	0 = Absent;1 = Present
34	Slumber party or sleepover	0 = Absent;1 = Present
35	Summoner is locked in room	0 = Absent;1 = Present
36*	Figure descends down stairs in Mirror	0 = Absent;1 = Present

^*Indicates characters coded as counterintuitive.

References

Ahrens

Locke

R. J.

Sawyer

(Producers) Stecenko

(Producer & Director) . (2005). The legend of Bloody Mary [Motion picture]. United States: MK4 Productions.

Armitage

(2006). ‘All about Mary’: Children’s use of the toilet ghost story as a mechanism for dealing with fear, but fear of what? Contemporary Legend, 9, 1–27.

Atran

Norenzayan

(2004). Religion’s evolutionary landscape: Counterintuition, commitment, compassion, communion. Behavioral and Brain Sciences, 27, 713–770.

Barker

Golin

(Producers) Rose

(Director) . (1992). Candyman [Motion picture]. United States: Propaganda Films.

Barrett

J. L.

(2000). Exploring the natural foundations of religion. Trends in Cognitive Sciences, 4, 29–34.

Barrett

J. L.

(2008). Coding and quantifying counterintuitiveness in religious concepts: Theoretical and methodological reflections. Method and Theory in the Study of Religion, 20, 308–338.

Barrett

J. L.

Burdett

E. R.

Porter

T. J.

(2009). Counterintuitiveness in folktales: Finding the cognitive optimum. Journal of Cognition and Culture, 9, 271–287.

Barrett

J. L.

Nyhof

M. A.

(2001). Spreading of non-natural concepts: The role if intuitive Conceptual structures in memory and transmission of cultural materials. Journal of Cognition and Culture, 1, 69–100.

Bartlett

F. C.

(1932). Remembering. Oxford, UK: Macmillan.

10.

Boyer

(1994). The naturalness of religious ideas: A cognitive theory of religion. Berkeley, CA: University of California Press.

11.

Boyer

Ramble

(2001). Cognitive templates for religious concepts: Cross-cultural evidence for recall of counter-intuitive representations. Cognitive Science, 25, 535–564.

12.

Brunvand

J. H.

(2000). The truth never stands in the way of a good story. Chicago: University of Illinois Press.

13.

Burns

(2009). The poems and songs of Robert Burns: The five foot shelf of classics (Vol. VI). New York, NY: Cossimo, Inc.

14.

Cameron

Dueck

Glazer

Harrison

Howsam

(Producers) Wilson

(Director) . (2007). Dead Mary [Motion picture]. Canada: 235 Films.

15.

Collard

Shennan

S. J.

Tehrani

J. J.

(2006). Branching, blending and the evolution of cultural similarities and differences among human populations. Evolution and Human Behavior, 27, 169–184.

16.

Dundes

(1998). Bloody Mary in the mirror: A ritual reflection of pre-pubescent anxiety. Western Folklore, 57, 119–135.

17.

Ellis

(2004). Lucifer ascending: The occult in folklore and popular culture. Lexington: University of Kentucky Press.

18.

Eriksson

Coultas

J. C.

(2012). The advantage of multiple cultural parents in the cultural transmission of stories. Evolution and Human Behaviour, 33, 251–259.

19.

Farris

J. S.

(1989). The retention index and homoplasy excess. Systematic Zoology, 38, 406–407.

20.

Felsenstein

(1989). PHYLIP - Phylogeny Inference Package (Version 3.2). Cladistics 5: 164–166. Retrieved June 3, 2012, from software homepage: http://evolution.gs.washington.edu/phylip.html)

21.

Fox Tree

J. E.

Weldon

M. S.

(2007). Retelling urban legends. American Journal of Psychology, 120, 459–476.

22.

Heath

Bell

Sternberg

(2001). Emotional selection in memes: The case of urban legends. Journal of Personality and Social Psychology, 81, 1028–1041.

23.

Hutton

(2001). Stations of the sun: A history of the ritual year in Britain. Oxford: Oxford University Press.

24.

Jones

W. H.

Kropf

L. L.

(1883). Magyar folk-lore and some parallels. The Folklore Journal, 1, 354–362.

25.

Klintberg

(1988). “Black Madame, come out!” On school children and spirits. Arv: Scandanavian Yearbook of Folklore, 44, 155–167.

26.

Knapp

(1976). One potato, two potato. . .. The secret education of American school children. New York, NY: W.W. Norton.

27.

Longlois

(1978). Mary Whales, I believe in you: Myth and ritual subdued. Indiana Folklore, 9, 5–33.

28.

Marques

Páez

Valencia

Vincze

(2006). Effects of group membership on the transmission of negative historical events. Psicologia Politica, 32, 79–103.

29.

Merrell

Phillips

Messer

(Producers) Lambert

(Director) (2005). Urban legends: Bloody Mary [Motion picture]. United States: NPP Productions.

30.

Mesoudi

(2011). Cultural evolution: How Darwinian theory can explain human culture and synthesize the social sciences. Chicago, IL: University of Chicago Press.

31.

Miller

T. W.

Tyler

(Producers) Valentine

(Director) (2006). Bloody Mary [Motion picture]. United States: KAT Pictures.

32.

Norenzayan

Atran

Faulkner

Schaller

(2006). Memory and mystery: The cultural selection of minimally counterintuitive narratives. Cognitive Science, 30, 531–553.

33.

Nunn

C. L.

Arnold

Matthews

Borgerhoff Mulder

(2010). Simulating trait evolution for cross-cultural comparison. Philosophical Transactions of the Royal Society B, 365, 3807–3819.

34.

Ryan

W. F.

Wigzell

(1992). Gullible girls and dreadful dreams. Zhukovski, Pushkin, and popular divination. The Slavonic and East European Review, 70, 647–669.

35.

Swofford

D. L.

(2003). PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sunderland, MA: Sinauer Associates. Retrieved June 3, 2012, from http://paup.csit.fsu.edu/

36.

Tangherlini

T. R.

(1990). “It happened not too far from here…”: A survey of legend theory and characterization, Western Folklore, 49, 371–390.

37.

Tucker

(2005). Ghosts in mirrors: Reflections of the self. Journal of American Folklore, 118, 186–203.

38.

Upal

M. A.

(2011). Memory, mystery and coherence: Does the presence of 2-3 counterintuitive concepts predict cultural success of a narrative? Journal of Cognition and Culture, 11, 23–48.

39.

Upal

M. A.

Gonce

L. O.

Tweney

R. D.

Slone

D. J.

(2007). Contextualising counterintuitiveness: How context affects comprehension and memorability of counterintuitive concepts. Cognitive Science, 31, 415–439.