In a series of recent papers, Nisbett and Cohen have argued that these regional differences reflect a "culture of honor" that arose in the frontier South and West because of their historical herding economies coupled with the lack of effective law enforcement. Livestock rustlers can quickly destroy a herder's livelihood. This danger would be quite palpable even if The Law were available since both thieves and booty could be long gone before a search party was mobilized. In frontier America, even this measure of law enforcement was often absent, and so herders were left to fend for themselves. In such an environment, herders deterred prospective thieves with a tough demeanor and an honor code that advertised one's family and property as sacred. Any violation of that sanctity would be met with quick and harsh retaliation.
Evidence for this "culture of honor" has been obtained through ethnographic, survey, and experimental investigations. First, culture of honor characteristics have appeared consistently in herding societies around the world (Schneider 1971). Second, attitude surveys have found that southern and western Americans do not condone violent responses to triggering events across the board. Rather, they differ from northern Americans only when honor is at stake (Cohen and Nisbett 1994). Furthermore, regional differences in homicide rates are restricted to violations of personal honor, such as barroom insults (Nisbett 1993). Third, southern and western businesses show more sympathy to job applicants who have criminal records when the crime involves a defense of honor than when it does not. Finally, when insulted, southern males respond with more anger than northern males and show stronger physiological signs of stress and aggression (e.g., increases in cortisol and testosterone levels; Cohen, Nisbett, Bowdle and Schwartz 1996).
Although the frontier history shared by the South and West could account for their similarities, there is some evidence that the South is significantly more violent than the West. For example, whereas southern and western juries in capital cases recommend the death penalty at equal rates, these sentences are actually carried out more frequently in the South. School discipline in the South is also more likely to involve corporal punishment (Cohen 1996). Cohen attributes these patterns to the long history of slavery and post- slavery racial subjugation in the South, where white dominance was maintained by both state sanctioned and vigilante violence. Cohen provided further evidence for this attribution by predicting and finding subpatterns in measures of violence within the South. In particular, he reasoned that states with greater levels of slavery should show higher levels of violence. He therefore compared the upper South and the lower South on various measures of violence. In comparison with slave states in the lower South like Alabama and Georgia, those in the upper South, such as Maryland and Delaware, had relatively small slave populations and remained in the Union during the Civil War. These historical differences are correlated with contemporary differences in institutionalized violence. Thus, death sentences have been carried out much more frequently in the lower South than in the upper South. Corporal punishment is also used in school discipline more often in the lower South.
In sum, extensive evidence has implicated important historical antecedents in southern and western violence. Still, it is not clear why these regional patterns persist when their precipitating conditions--slavery and/or a precarious herding lifestyle in a dangerous frontier--no longer exist. In their work on the culture of honor, Cohen and Nisbett (1997) suggest that various institutional forces may help to maintain this culture and its violent consequences. For example, Cohen and Nisbett found that college newspapers in southern and western states showed more sympathy than northern newspapers to crimes committed in defense of honor (e.g., a stabbing provoked by an insult to one's sister). No regional differences were found in descriptions of a robbery, which did not involve an honor violation. The survival of certain "culture of honor" patterns in the legal systems of southern and western states might also reinforce congruent attitudes. For example, northern states generally have "retreat rules," whereby a person must first attempt to retreat from a potential assailant before giving a more violent defense. Legal codes in southern and western states lack such rules, and historical analysis indicates that residents viewed them as humiliating affronts to personal honor (Cohen 1996). Current residents of southern and western states could assimilate certain attitudes toward violence through exposure to such legal codes and other institutions established during the frontier period and, due to their conservative nature, surviving its demise.
In this paper, I will test whether certain linguistic patterns exist that reflect the prominence of violence in the South and West, and might help to promote the continued acceptability of violence in these areas. In particular, I will focus on regional differences in the prevalence of proper names that have violent connotations, such as names of locations (e.g., "Murderer's Creek") and businesses (e.g., "Gunbarrel Liquors"). Naming is a very important linguistic behavior personally and socially. Names testify to family and cultural identity, reinforce the values associated with those identifications, and advertise them publicly (e.g., Fischer 1989). To the extent that violence is more salient and positively valued in the South and West than in the North, then names with violent connotations should be more prevalent in those regions of the United States.
Many places throughout the United States have names based on proper nouns, such as people of local, national, or international importance. But proper names have not been the sole source of American place names as many sites have been dubbed with common nouns, such as "Ash Grove," NY and "Lobster Cove," ME. In addition to choosing locally important (and innocuous) fruits and animals for place names, however, Americans have also named some sites after violent objects ("Guntown," MS), actions ("Murder Hill," NY), and persons ("Cutthroat Gulch," CA). Given the history of violence in the South and West, such names should be more prevalent there than in the North. Regional differences should be less apparent for place names that lack violent connotations.
Regional definitions Southern, western, and northern regions of the continental United States were defined according to Cohen and Nisbett (1997), who in turn relied on census classifications. The sixteen southern states were Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia (Census Divisions 5, 6, and 7). The eleven western states were Arizona, California, Colorado, Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, and Wyoming (Census Divisions 8 and 9). The remaining twenty-one states were considered northern.
Place name source. The United States Geological Survey's on-line database of place names was the source for this study's data (Internet address: www-nmd.usgs.gov/www/gnis). The database, known as the Geographic Names Information System (GNIS; see McArthur 1995 and Payne 1995, for further information), returned all place names in its registry that began with the keywords described in the next section, along with the state and county in which each place was located, and the type of feature so identified (e.g., lake, summit, park, populated place like a city or town). Unfortunately, the current version of the GNIS search engine does not return places that contain a keyword in non-initial position. However, language comprehension research indicates that words at the beginning of phrases are particularly memorable (Gernsbacher and Hargreaves 1988), and so it seems reasonable to focus on place names that start with violent words.
Violent keywords and controls. The following procedure was used to select keywords with violent meanings. First, I searched the electronic version of the American Heritage Dictionary for all words that contained "weapon," "firearm," "armed," "combat," "conflict," "kill," "killer," "murder," "murderer," "criminal," "violent," or "violence" in their definitions. These words were selected because they appeared in the definitions of many salient violent words, such as "gun," "bullet," and "war." This search returned 638 words, with phrases like "foul play" and proper nouns excluded. However, many of these items had multiple meanings, some of which were nonviolent. For example, "cutthroat" can refer to a type of trout as well as a murderer. "Sight" can denote the sense of vision or an aiming device on a firearm. Hence, it is not clear that place names beginning with such words derive from violent meanings. To verify that each place name draws upon the violent meaning of its initial word would be a prodigious undertaking. However, a safeguard was imposed to reduce the ambiguity problem. In particular, words were only considered violent if their primary meaning in The American Heritage Dictionary contained one or more of the words used to obtain the original list (e.g., "weapon"). Thus, "cutthroat" was included because its murderer meaning is primary. "Sight," on the other hand, was excluded because its dominant visual sense meaning did not involve violence. This definitional criterion led to the removal of 439 words from the list.1 A representative sample of the remaining 199 words was drawn for place name analysis. Words were selected quasirandomly in that the odds of selection were proportional to word frequency, with frequency values obtained from Francis and Kucera (1984). The frequency bias was used to increase the chances that a large number of places would be involved in the analysis. Words were submitted to the GNIS search engine in order of selection until twenty successful searches were obtained. A search was considered successful if the GNIS returned at least ten place names within the 48 contiguous states that began with the target keyword. To meet this constraint, 34 words were actually submitted to the GNIS search engine, but accused, assassin, attack, combat, conflict, corps, destroy, duel, fling, invasion, rebellion, strife, tripwire, and violent produced fewer than ten matches. The 20 words that met the frequency requirement are listed in Appendix 1. They include names for weapons (e.g., "gun," "bomb"), violent persons (e.g., "cutthroat," "outlaw"), and events (e.g., "murder," "war"). To extend the range of the sample, the results included morphological variants for each keyword. For example, the tallies for "gun" included place names beginning with "gunpowder," "gunbarrel," and "guntown" among others. Similarly, "warpath," "warrior," and related words were included in the counts for "war."
Some of the selected keywords, such as "spear," could pose interpretive problems because they have surname uses. There are three reasons why such ambiguities do not appear serious, however. First, if surnames with violent connotations are distributed uniformly throughout American place names, then regional differences in common noun usage will be diluted. Hence, the surname ambiguity would only underestimate regional differences. Second, if the violent connotations of some surnames make them attractive place names for southerners and westerners, then their use provides another reflection of regional values. Third, suppose many place names beginning with "spear" derive historically from a personal name. Most Americans who currently live in or near these sites would not be aware of the historical background of the place names, and so would tend to interpret them in terms of their common violent meanings.
To test adequately the hypothesis that violent place names cluster in the South and West, control words are needed to rule out artifactual accounts of the results. For instance, there are 27 southern and western states in the analysis, compared with 21 northern states. Furthermore, the former states comprise a much larger surface area than the latter, and hence have more potential places to name. Any name type might then be more common in the South and West than in the North. The following procedure was therefore used to obtain a set of 20 control words. First, proper names like "Franklin" were excluded from consideration as control words since the violent word set consisted of common nouns, verbs, and adjectives rather than names. The control set should have a similar makeup. Next, to increase the ease of finding common words that serve as place names, a dictionary of American place names was consulted (Stewart 1970). This dictionary was sampled by randomly selecting a page and then submitting the first common word entry to the GNIS database. If the names of ten or more locations began with that word, it was included in the control set. This random selection process continued until 20 control words were obtained. Twenty-four words needed to be sampled to obtain 20 items that met the frequency requirement. The four words that did not meet the frequency criterion were abalone, bauxite, curiosity, and emblem. The control words are listed in Appendix 1. As can be seen, the meanings of the control words are clearly less fierce than those of the violent words.
The number of place names that began with each keyword was tallied separately for southern, western, and northern states. The proportion of such place names that occurred in the South and West was then calculated for each keyword. For example, there were 296 place names that began with "gun," with 109 located in southern states, 121 in western states, and 66 in northern states. Hence, 78% of these "gun" place names appeared in southern and western states. Similar proportions were calculated for the remaining 19 violent words and the 20 control words, and the means for the two word classes compared using t-tests.
As predicted, the mean proportion of place names with violent words was higher in the South and West than in the North. On average, 80% of places beginning with violent names were in the South and West. In contrast, only 66% of places beginning with the non-violent control words were located in these regions (t(38) = 3.21, p < .005). This pattern appeared when the West and South were compared separately with the North. When names just in the South and North were considered, 63% of violent names appeared in the South compared with 44% of the non-violent names (t(38) = 2.78, p < .01). Similarly, the ratio of violent names in the West to those in the combined West plus North was 66%. The corresponding ratio for the control words was only 51% (t(38 = 2.39, p < .03). The West and South, on the other hand, did not differ significantly in the incidence of places named with violent words compared with control words as 48% of the violent names and 41% of the nonviolent names appeared in the South (t(38) = 0.77, p > .40). However, the pattern is consistent with the stronger history of violence in the South (Cohen 1996).
Two potential problems with the results presented so far need to be addressed. First, since the states were not considered separately, the regional differences could be due to a couple of atypical states. For instance, the differences could be carried by one or two northern states with an extremely low incidence of violent place names. Second, the initial statistical tests used the keyword as the unit of analysis, and so treated each word equally even though they varied greatly in the number of place names that began with them. For instance, 541 place names began with "battle" but only 17 began with "cutthroat." Since "battle" and "cutthroat" were treated equally in the statistics, place names that began with the low frequency word were effectively given more weight. The statistical analyses would therefore exaggerate regional effects if the high frequency violent words were distributed uniformly across the United States, but the low frequency words were concentrated in the South and West.
To address these two problems, a second set of statistical tests was performed with the state as the unit of analysis. Each state was scored for the proportion of violent place names among the total place names recorded for that state. Each individual place name was therefore weighed equally in this scoring system, thus giving words like "battle," which appeared in many place names, a greater impact in this analysis than in the first one. The resulting mean proportions of violent names were .51, .44, and .32 for the South, West, and North respectively. All pairwise comparisons among these means were significant (South vs. West, t(25) = 2.43; South vs. North, t(35) = 6.32; West vs. North, t(30) = 3.41; all ps < .03). Thus, the results for the keyword analysis were replicated in this state analysis, with the additional finding that violent place names were significantly more common in southern than western states. This pattern is consistent with data indicating more positive attitudes toward violence in the South (Cohen 1996).
In most circumstances, statistical significance is used to generalize from a sample to a population. However, that induction is meaningless in this case since the entire population of relevant states (i.e., those in the continental US) was examined. Still, the statistical significance does refute the possibility that the results from the first analysis were only due to a few atypical states. This is further confirmed when the states are ranked according to proportion of violent place names. Not one northern state appeared among the ten states with the highest violence proportions whereas nine such states appeared among the ten with the lowest proportions. Furthermore, only two northern states appeared among the twenty states with the highest proportions whereas sixteen appeared among the twenty with the lowest proportions.
As one further illustration of regional differences in violent place names, consider the distributions of place names beginning with the opposites "war" and "peace." In the North, "peace" places outnumber "war" places by two to one in the North (127 to 62). In contrast, "war" places outnumber "peace" places by almost two to one in the South (193 to 113) and three to one in the West (95 to 34).
Study 1 documented the greater prevalence of violent words in southern and western place names than in northern place names. This difference is consistent with the historically higher levels of violence in the South and West, and these fossil remnants of that history may help to maintain current regional differences in violent attitudes and behavior. Such effects might be amplified if violent words in more recent name coinages also cluster in the South and West. Study 2 will test this possibility in the area of business names.
Regional Definitions. Northern, southern, and western states were defined as in Study 1.
Business Name Source. DeLorme's electronic telephone directory known as Phone Search USA (1997, Version 3.0) was used to find businesses in the continental United States that began with certain keywords. Like the GNIS source for place names in Study 1, the program only returned business names that began with the keywords.
Violent Keywords and Controls. A new set of violent words was randomly drawn from the list described in Study 1. Since selection was weighted by frequency, many of the keywords from study 1 were also chosen here. Keywords were submitted to the electronic directory in order of selection until twenty successful searches were obtained. Successful searches returned at least ten businesses that began with the target word or a morphological variant, such as "gunsmoke" for the keyword "gun" or "cannonball" for the keyword "cannon." Thirty-two words had to be submitted to the directory search to meet this frequency requirement. The words that failed the constraint were accused, artillery, assault, backlash, conscript, destroy, invasion, ordnance, prosecutor, robbery, shrapnel, and weapon.
Twenty nonviolent control words were drawn from the Battig and Montague (1968) category dominance norms. These norms were generated by presenting subjects with a category name, such as "fruit," and then asking them to list as many category members as they could recall in 30 seconds. Category dominance was then measured by the number of subjects who listed a particular category member. I obtained nonviolent control words from these norms by randomly selecting twenty categories and, within each of these categories, one of the five most dominant members. Business names that began with each of these control words were then extracted from the directory. Appendix 2 lists the violent and control keywords for this study.
An initial examination of the search results revealed many businesses that dealt in violent products or services, such as firearms and self-defense training. It is not surprising that the name for a gun shop begins with the word "gun." Any regional differences in the prevalence of such shops could create corresponding differences in the frequency with which violent words appear in business names. However, such differences would not reflect direct effects of culture on name choice, but only indirect effects mediated by correlations between culture and business proclivities. It would be more interesting to find violent words used to name businesses like restaurants or automobile repair shops that do not have an intrinsic connection with violence. Name choices here could be attributed more easily to the overall cultural milieu. Violent businesses were therefore excluded from the analyses. These cases could be identified easily and objectively because the directory coded each listing for business type. Appendix 3 lists the exclusions. Some listings did not have a specific code, and were accordingly labeled with a "0." The business type could be inferred from the name for many of these cases (e.g., "Killer Chicken CafÚ"), but just to be conservative, businesses coded with "0" were also dropped from the analyses.
Another potential confound in the data concerns the lack of independence between business and place names. In particular, many businesses are named after the city or town in which they are located (e.g., "War Taxi" in War, WV and "Rifle Realty" in Rifle, CO). The presence of such cases in the data could make it seem as though southern and western businesses are named more often with violent words. However, such effects would in reality be driven by the greater frequency of violent place names in the South and West rather than any additional inclination to use violent names by current residents of these regions. I therefore excluded businesses from the analyses if their names began with the same word as the city or town in which they resided. This restriction was imposed on the businesses in both the violent condition and the nonviolent control condition. The latter needed to be included because the relative scarcity of violent place names in the North could artificially dilute the presence of violent business names in this region.
The number of businesses that began with each keyword was tabulated separately for southern, western, and northern states. Each keyword then received a score corresponding with the proportion of hits that occurred in the southern and western states. The means for the violent and nonviolent words were then compared. As predicted, business names that began with violent words came from the South and West more than names that began with the nonviolent control words. On average, 68% of the violent names came from southern and western businesses compared with only 53% of the nonviolent names (t(38) = 4.07, p < .0005). The relative scarcity of violent names in northern businesses appeared in separate comparisons with the South and West. Among southern and northern businesses, 58% of those with violent names were located in the South, but only 39% of those with nonviolent names (t(38) = 4.06, p < .0005). Among western and northern businesses, 45% of those with violent names were located in the West compared with only 34% of those with nonviolent names (t(38) = 2.75, p < .01). Relative to the nonviolent control words, the South and West did not differ significantly in the prevalence of violent business names as 61% of violent names and 56% of nonviolent names appeared in southern businesses (t(38) = 1.11, p > .20). Once again, however, the trend is consistent with other evidence showing a stronger culture of violence in the South (Cohen 1996).
As in the place name study, statistical tests were also conducted with the state as the unit of analysis. Each state was scored for the proportion of businesses with violent names among the total businesses returned in the directory search. The mean proportions were .14, .13, and .08 for the southern, western, and northern states. Both the southern and western businesses contained a significantly higher proportion of violent names than the northern businesses (South vs. North: t(35) = 7.12; West vs. North: t(30) = 3.60, ps < .005). The values for the South and West were not significantly different (t(25) = 0.55). When the states were ranked according to proportion of businesses with violent names, not one northern state appeared among the top ten and only one appeared among the top twenty. In contrast, nine northern states were among the bottom ten and fifteen among the bottom twenty.
If so inclined, a family in Alabama could have their television serviced at Warrior Electronics, their dog housed at Gunsmoke Kennels, their home addition built by Bullet Construction, and their children taught at Battleground School. A Texan could be born in Gun Barrel City, pray at Battle Ax Church, fish at Bullet Creek, dine at Shotguns Bar BQ, work at Outlaw Avionics, and be interred in Battle Creek Cemetery. Of course, northerners can also encounter places or businesses with violent names, such as Bloody Pond, New York and Shotgun Willie's Saloon in Massachusetts. In a nation like the United States, one does not have to travel far to see violence highlighted. However, such opportunities are rarer in the North than in the South and West.
There is as yet no evidence that these linguistic patterns affect attitudes or behaviors involving violence. In particular, we cannot say whether the predominance of violent names in the South and West helps to maintain the greater acceptability of violence in these regions. However, there are many psychological mechanisms that could produce such effects. For example, the more frequently an object is encountered, the more it is liked in general (Zajonc 1968). The greater frequency of violent words in southern and western names could therefore increase the positive evaluations given to the objects and actions that those words denote. More generally, high frequency conveys many perceptual and cognitive advantages. Thus, high frequency words are easier to identify (e.g., Forster and Chambers 1973) and recall (e.g., Rubin and Friendly, 1986) than low frequency words. While driving down a road crowded with businesses and their banners, hungry southerners and westerners should find it easier than northerners to identify or remember a restaurant with a violent name. More interestingly, research suggests that the attractiveness of an object increases when it is easier to identify (Bornstein and D'Agostino 1992). The greater frequency of violent names in the South and West might not only make such names and their meanings easier to recognize by residents of those regions, but, as a direct consequence, also more attractive or desirable.
In sum, the greater environmental availability of violent names in the South and West will make them more cognitively available to residents of those regions than to northerners. The research discussed above suggests that this increased availability could affect attitudes toward violence. But could violent behavior also be influenced? There is evidence that increasing the environmental and cognitive availability of violence can indeed stimulate violent actions (see Berkowitz 1974 and 1984 for reviews). For example, Berkowitz and LaPage (1967) placed undergraduates in a situation where they could use mild but annoying electric shocks to punish a fellow student for mistakes on a laboratory task. More shocks were administered when the laboratory contained some salient violent objects, such as a gun, than when it contained neutral objects, such as a badminton racquet. Of more relevance to the present paper, similar effects occurred when subjects were merely exposed to words with violent meanings (Loew 1967; Parke, Ewall, and Slaby 1972; Turner and Layton 1976). For example, Turner and Layton (1976) asked their subjects to first learn a set of violent or neutral words. Subjects exposed to the violent words later gave more intense shocks to a fellow participant who made mistakes on a laboratory task. Such experiments suggest that mere exposure to violent words could increase predispositions to violence, perhaps by activating violent thoughts and schemas with which those words are associated (Berkowitz 1974; Anderson, Benjamin and Bartholow 1998). Such effects might be amplified where violent words are particular frequent, as in the South and West.
Violent words and concepts might also inherit some acceptability from the socially positive entities with which they are associated through naming, such as schools, churches, and respected businesses. One possible mechanism for such inheritance could involve a desire for cognitive consistency (see Sabini 1992 for a review of this large literature). A person who regularly attends church services presumably has positive attitudes toward religion. If that person also had negative or neutral attitudes toward guns, an apparent conflict would arise while praying in "Rifle Range Church." Cognitive consistency could be achieved by lowering one's opinion of religion or raising one's opinion of firearms, and history suggests that the church militant generally triumphs over atheistic pacifism. Of course, any positive views toward violence might also generalize to businesses with which they are associated. Thus, a business in the South and West might be regarded more favorably if it contained violent words in its name. Smaller or opposite effects should be found with similar businesses in the North.
In sum, regional differences in American values concerning violence predict corresponding regional differences in name usage. On many measures, Americans from southern and western states view violence more positively and constructively than Americans from northern states. The research presented here found that place and business names in the South and West are more likely than those in the North to begin with violent words. Southerners and westerners who hear, read, say, or write these names are therefore given repeated reminders of their regional values. It remains to be seen whether such exposures reinforce those values, but given the regular association of violent words with positive objects in the South and West, it is doubtful that such words erode them. After all, how profane can guns, pistols, rifles, bullets, etc. be when they are repeatedly linked to the sacred in places like "Gunpowder Church" and "Bullets Chapel?"
Anderson, C.A., A.J. Benjamin, and B.D. Bartholow. 1998. "Does the Gun Pull the Trigger? Automatic Priming Effects of Weapon Pictures and Weapon Names." Psychological Science 9: 308-314.
Battig, W.F., and W.E. Montague. 1969. "Category Norms for Verbal Items in 56 Categories: A Replication of the Connecticut Category Norms." Journal of Experimental Psychology 80: (3, pt. 2).
Baron, L., and M.A. Straus. 1988. "Cultural and Economic Sources of Homicide in the United States." Sociological Quarterly 29: 371-392.
Berkowitz, L. 1974. "Some Determinants of Impulsive Aggression: Role of Mediated Associations with Reinforcement for Aggression." Psychological Review 81: 165-176.
Berkowitz, L. 1984. "Some Effects of Thoughts on Anti- and Prosocial Influences of Media Events: A Cognitive- Neoassociation Analysis." Psychological Bulletin 95: 410-427.
Berkowitz, L., and A. LaPage. 1967. "Weapons as Aggression- Eliciting Stimuli." Journal of Personality and Social Psychology 7: 202-207.
Bornstein, R. F., and P.A. D'Agostino. 1992. "Stimulus Recognition and the Mere Exposure Effect." Journal of Personality and Social Psychology 63: 545-552.
Cohen, D. 1996. "Law, Social Policy, and Violence: The Impact of Regional Cultures." Journal of Personality and Social Psychology 70: 961-978.
Cohen, D., and R.E. Nisbett. 1994. "Self-Protection and the Culture of Honor: Explaining Southern Violence." Personality and Social Psychology Bulletin 20: 551-567.
Cohen, D., and R.E. Nisbett. 1997. "Field Experiments Examining The Culture Of Honor: The Role Of Institutions In Perpetuating Norms About Violence." Personality and Social Psychology Bulletin 23: 1188-1199.
Cohen, D., R.E. Nisbett, B.F. Bowdle, and N. Schwarz. 1996. "Insult, Aggression, and the Southern Culture Of Honor: An Experimental Ethnography." Journal of Personality and Social Psychology 70: 945-960.
Fischer, D. H. 1989. "Albion's Seed." New York: Oxford University Press.
Forster, K. L., & S. M. Chambers. 1993. "Lexical Access and Naming Time." Journal of Verbal Learning and Verbal Behavior 18: 1-20.
Francis, W. N., & H. Kucera. 1982. "Frequency Analysis of English Usage: Lexicon and Grammar." Boston: Houghton- Mifflin.
Gernsbacher, M. A., and D.J. Hargreaves. 1988. "Accessing Sentence Participants: The Advantage of First Mention." Journal of Memory and Language 27: 699-717.
Lee, R. S. 1995. "Regional Subcultures as Revealed by Magazine Circulation Patterns." Cross-Cultural Research 29: 91-120.
Loew, C. A. 1967. "Acquisition of a Hostile Attitude and its Relationship to Aggressive Behavior." Journal of Personality and Social Psychology 5: 335-341.
McArthur, L. L. 1995. "The GNIS and the PC: Two Tools for Today's Toponymic Research." Names 43: 245-254.
Nisbett, R. E. 1993. "Violence and U.S. Regional Culture." American Psychologist 48: 441-449.
Parke, R. D., W. Ewall, and R. C. Slaby. 1972. "Hostile and Helpful Verbalizations as Regulators of Non-Verbal Aggression." Journal of Personality and Social Psychology 23: 243-248.
Payne, R. L. 1995. "Development and Implementation of the National Geographic Names Database." Names 43: 307-314.
Rubin, D. C., & M. Friendly. 1986. "Predicting which Words Get Recalled: Measures of Free Recall, Availability, Goodness, Emotionality, and Pronunciability for 925 Nouns." Memory and Cognition 14: 79-94.
Sabini, J. 1992. Social psychology. New York: Norton.
Schneider, J. 1971. "Of Vigilance and Virgins: Honor, Shame, and Access to Resources in Mediterranean Societies." Ethnology 10: 1-24.
Stewart, G. R. 1970. American Place-Names. New York: Oxford University Press.
Turner, C. W., & J. F. Layton. 1976. "Verbal Imagery and Connotation as Memory-Induced Mediators of Aggressive Behavior." Journal of Personality and Social Psychology 33: 755-763.
Zajonc, R. B. 1968. "Attitudinal Effects of Mere Exposure." Journal of Personality and Social Psychology Monograph Supplement 9: 1-27.
1. The seriousness of the ambiguity problem is also lessened by the fact that alternative meanings to the violent keywords vary idiosyncratically from word to word. The one aspect of meaning that they share is their violent connotation. Hence, the use of multiple violent keywords and appropriate statistical tests can minimize the possibility that the results are spurious.
Violent Keywords: battle, blade, bomb, bullet, cutthroat, fight, fury, gun, kill, military, murder, outlaw, pistol, poison, rifle, shoot, spear, strike, trigger, war
Nonviolent keywords: ash, chalk, cheese, eel, emerald, glass, lizard, lobster, mitten, moose, omega, pea, square, strawberry, temperance, turnip, twilight, vine, wagon, wheat
Violent Keywords: attack, battle, bayonet, bomb, bullet, cannon, explosion, fight, gun, kill, military, missile, murder, outlaw, overcome, pistol, rifle, shoot, smash, war
Nonviolent Keyowords (and categories): apple (fruit), bean (vegetable), copper (chemical element), dog (four-legged animal), doll (toy), green (color), head (body part), hill (natural earth formation), maple (tree), sapphire (precious stone), saw (tool), shirt (clothing), sparrow (bird), table (furniture), tent (dwelling), trout (fish), tulip (flower), vanilla (flavoring), water (nonalcoholic beverage), wool (type of cloth)