Framing violence: A content/discourse analysis of representations of violence in Bahamian newspapers


Material Information

Framing violence: A content/discourse analysis of representations of violence in Bahamian newspapers
Physical Description:
15 p.
Oenbring, Raymond.
College of The Bahamas
Place of Publication:
Nassau, Bahamas
Publication Date:


Subjects / Keywords:
Violence--Bahamas--Newspapers.   ( lcsh )


General Note:
This study provides a discourse regarding representations of violence in Bahamian newspapers.

Record Information

Source Institution:
College of The Bahamas
Holding Location:
College of The Bahamas
Rights Management:
All rights reserved by the source institution.
System ID:

This item is only available as the following downloads:

Full Text


Draft: Violence Symposium 3rd November 2011 Page | 1 Framing Violence: A Content/Discourse Anal ysis of Representations of Violence in Bahamian Newspapers Raymond Oenbring School of English Studies Abstract This project is the first schol arly study of representations of violence in Bahamian newspapers. Using the methods of discourse analysis, content an alysis, and corpus linguistics, the study compares the content of a corpus of violent crime articles from the Nassau Tribune and Guardian with similar corpora of violent crime articl es collected from the leading establishment newspapers of Jamaica, Trinidad, and South Flor ida. Features searched for include N-grams (frequently reoccurring strings of words) and keyw ords (terms occurring at a higher rate in one body of text versus another). The study finds th at, in general, newspa per articles reporting on violent crime in the Bahamas tend to rely heavily upon stock phrases and also, following Standard Bahamian English in general, rely heavily upon passive voice expressions, expressions that may remove the agency of the actor s involved. The study finds no evidence of sensationalism of violence in mainstream Bahami an newspaper articles, at least on the linguistic level.


Draft: Violence Symposium 3rd November 2011 Page | 2 Introduction As Bahamian journalist Noelle Nicolls notes in her recent commentary piece in The Tribune “How far can the media go in search of trut h?” (2010)lamenting the state of journalism in the Bahamas, “The representations of reality that bear themselves on the pages of newspapers and in the images of a television broadcast are si mply constructs of reality; they are angles. Understanding them as anything different is to be drawn into an illusion; to conflate the opinions of a few with absolute truth.” In this pithy quot e Nicolls affirms in the Bahamian context several broadly held assumptions among contemporary academics in the social sciences and humanities around the globe: specifically, that language practices — including t hose of the media—in effect create reality for mu ch of the population. Indeed, one of the major contentions of curre nt research in the humanities and social science falling under the multid isciplinary banners of cultural studies and postmodernism is that the available language practices of a given society constrain how individuals in that society think about the world (see, for example, Mills [1997] and Jaworski and Coupland [1999]). That is to say, language constructs conventional wisdom. To categorize and critique the language practices of socially important practices of representation, includi ng the discourse of the media, scholars in the humanities and social sciences have developed numerous sets of methods, including discourse analysis, cont ent analysis, narrative analysis, corpus linguistic analysis — several of which are put to use in the current study. Discourse analysis is a set of methods orig inally stemming from linguistics but now put to use in many areas of the humanities and social sciences used by researchers to interrogate the reigning practices of language in a given area. Journals such as Critical Approaches to Discourse Analysis across the Disciplines have developed so that research ers can critically interrogate


Draft: Violence Symposium 3rd November 2011 Page | 3 socially important discourses such as mainstre am media. However, to date there no discourse analytic studies of language prac tices in contemporary Bahamian media. A similar, yet more quantitatively-focused set of methods for asking similar questions used more frequently in communication studies is conten t analysis (see, for example, Major and Atwood [2004] and Krippendorff [2004]). The current study is the first scholarly study of Bahamian media using the methods of discourse analysis, content analysis and corpus linguistics. The goa l of this study is to analyse, categorize and critique several of the most im portant and unique features of Bahamian media discourse about violent crime. However, to ma ke project more manageable, I have limited my analysis to newspaper articles reporting on acts of violence.In particular, my goal has been to flesh out the conventional terms, and stock expr essions in Bahamian newspaper writing about violence.In a less theoretical ve in, this study is a contribution to recent discussions in the Bahamas regarding whether or not Bahamian me dia outlets sensationali se violence, thereby potentially damaging the tourism sector (see, fo r example, the late 2009 debate regarding the reporting of violence in Bahamian media after the holdup of 18 tourists on New Providence in November, 2009 [e.g., “Bahamas Can’t Hide Crime Problem”]). Methods For the study, I collected corpora of 50 arti cles of reporting on violent crime using a weapon from various newspapers in several different locations around the English-speaking Caribbean. After initial analysis of the Bahamas corpus collected in spring 2010, I determined that it would be useful to deve lop similar corpora of newspaper writing on violence for several different locations around the English-speaking Caribbean for comparison. Accordingly, corpora


Draft: Violence Symposium 3rd November 2011 Page | 4 for South Florida, Jamaica, and Trinidad were collected during summer 2011. All of the articles were collected from the websites of the respective newspapers, and were save d in .txt format to facilitate analysis. The newspapers selected for each region we re mainstream, established publications within their respective regi ons –what are often called newspapers of record In addition to being established, the papers also had to do substantial reporting of local incidents of violent crime. (This precluded more national news papers in the United States [e.g., The New York Times, the Wall Street Journal ]). Tabloid-style publications such as the Punch were avoided. (Nonetheless, I recognize that the Punch and similar publications in other countries may have much greater effects on discourse about violence among the genera l populations of their re spective countries.) Articles selected for the corpora included prim ary reporting on violent in cidents, court cases stemming from violent incidents, and commentary pi eces about violence in the local regions. To be included the resp ective corpora, articles must involve either threatened or real acts of violence — including both actual cases, and mere threats of physical harm involving weapons (that is to say, armed robberies and related cr imes involving only the threat of physical violence were still treated as acts of violence). Ar ticles on violent actions from na tions and regions other than the local area of the respective newspapers, often reprinted from wire services such as the Associated Press, were also in cluded and treated the same as other articles. Oftentimes, several articles reporting on the same inci dent were included in the corpora. It should also be noted that articles reprinted on the newspapers’ websites may not reflect exactly the exact patterns of articles printed in each of the newspapers’ print versions. What’s more, the location of the articles in the original print versi ons was not taken into consideration.


Draft: Violence Symposium 3rd November 2011 Page | 5 For the Bahamian corpus, most of the articles were selected from The Tribune, with a handful of articles coming from each of The NassauGuardian, and The Freeport News. For Jamaica, articles were selected from the websites of the Jamaica Gleaner, and the Jamaica Observer. For South Florida (chosen to represent th e United States due to its geographical proximity and cultural bonds to the Caribbean proper), articles were selected from the Miami Herald and South Florida Sun-Sentinel. For the Trinidad corpus, arti cles were selected from the Trinidad Guardian and Express. For ease, this study shall refe r to the corpora as the BVCC (Bahamas Violent Crime Corpus), the SFVCC (S outh Florida Violent Crime Corpus), the JVCC (Jamaica Violent Crime Corpus), and the TVCC (Trinidad Violent Crime Corpus). Once each corpus was collected, a content anal ysis was performed in which were sorted based on the types of incidents reported on. Artic les were placed into one or more of the following categories: firearm incide nts involving death (including in cidents in which the death is a suspect at the hands of police); knife/machet e incidents involving death; cause of death unknown or unreported; incidents of death with other cause of d eath; firearm incidents without death; knife/machete incidents w ithout death; assaults without death involving other weapon or not identified weapon; commentaries on violen ce (including op-eds, social ramifications commentaries, and aftermath stories as part or all of the pi ece); and, finally, pieces in which is robbery identified as motive. For the corpus linguistic anal ysis, each corpus was search ed using corpus linguistic software packages (specifically, Wordsmith a nd AntConc) for N-grams and keywords. Broadly stated, the goal of corpus linguistic analysis is to search for patterns in a group of text (the corpus) using computer programs to speed up the tedious process of tallying features, and to uncover patterns that would not necessarily be cl ear to the naked eye. N-grams are frequently


Draft: Violence Symposium 3rd November 2011 Page | 6 reoccurring strings of word s with the corpus. Particular atten tion was paid to strings four words or longer (e.g., 4-grams and 5-grams). The goal of n-gram analysis is to tease out commonly used stock expressions within a particular corpus. Conve rsely, keywords are words that are determined to occur at a significantly higher rate in one corpus in comparison to another corpus that is used as a baseline. For the purposes of this study, the max p value was set to 0.000001. For a more detailed explanation of corpus li nguistic methods see Oenbring (2010), Fielding and Oenbring, et al. (2011), and/or Bi ber, Conrad, and Reppen (1998). Results The resultsof the content analysis suggest that incidents of ar med robbery are more frequently reported in the BVCC than any of the other corpora (see, for example, the comparatively higher rates of robbery being indi cated and firearm incidents not involving death in the BVCC). Note that this should not be read as a difference in patterns of incidents of violent crime, only a as a difference in the reporting of violent crime. A possi ble explanation for this difference could be as simple as the fact that the Bahamas has a smaller population than any of the other respective areas (and a lower murder count than both Jamaica and Trinidad). Thus, one might expect armed robbery to be a more newswo rthy event in the Bahamas. Nonetheless, the difference seems noteworthy. At the time the JV CC was collected, Jamaica was experiencing a horrific spike in beheadings. Thus, we should view the 50% knife/machete incidents involvingdeath figure as an outlier.


Draft: Violence Symposium 3rd November 2011 Page | 7 Table 1: Content Analysis Bahamas South Florida Jamaica Trinidad Firearm incident without death 35%26% 0%8% Robbery identified 30%14% 4%20% Firearm incident involving death 26%40% 38%44% Knife/machete incident without death 17%0% 4%8% COD unknown or unreported 11%10% 4%10% Assault without death other weapon or not identified 7%0% 2%4% Commentary 6%2% 12%2% Knife/machete incident involving death 2%16% 50%18% Other COD 6%6% 4%6% Stock Expressions From a stylistic perspective, perhaps the most interesting feature of the BVCC is the large number of stock expressions, expres sions not used at as high of a rate in the other violent crime corpora (see table 2). As a w hole, the high rate of stock expr essions suggests that Bahamian newspapers demonstrate a preferen ce for straight, uneditorialized reporting of violent crime. One might go so far as to describe the bulk of Bahamian newspaper writing on the subject of violent crime as staid, perfunctor y, or even tending toward boilerp late. To explain this pattern, one might look to the Bahamian esteem for formal ity and protocol, a cult ural pattern noted in many areas of Bahamian life and culture.


Draft: Violence Symposium 3rd November 2011 Page | 8 Table 2: Stock Expressions in Baha mas Violent Crime Corpus (4-grams) Rank Frequency N-Gram 1 20 was taken to hospital 2 15 armed with a handgun 3 14 taken to hospital by 4 11 a year old man 5 10 an undetermined amount of 6 10 not required to enter 7 9 a group of men 8 9 man armed with a 9 9 required to enter a 10 9 to enter a plea 11 9 undetermined amount of cash 12 8 a man armed with 13 8 a plea to the 14 8 and a year old 15 8 enter a plea to 16 8 is listed in stable 17 8 listed in stable condition 18 8 of an armed robbery 19 7 a small amount of 20 7 an armed robbery at


Draft: Violence Symposium 3rd November 2011 Page | 9 Rank Frequency N-Gram 21 7 approached by a man 22 7 Her Majesty s Prison 23 7 in an unknown direction 24 7 officer Sergeant Chrislyn Skippings 25 7 the Central Detective Unit 26 7 to Her Majesty s 27 7 was approached by a 28 7 were not required to 29 6 a search warrant on 30 6 A year old man 31 6 arrived at the scene 32 6 by a man armed 33 6 caused the death of 34 6 executed a search warrant 35 6 in the parking lot 36 6 It is alleged that 37 6 of an undetermined amount 38 6 pleaded not guilty to 39 6 police received information of


Draft: Violence Symposium 3rd November 2011 Page | 10 As Oenbring (2010) notes, Standard Bahamian English, that is the variant of Standard English used in formal situa tions, including in news reportin g, often displays overly formal preferences of usage (e.g., the legalistic persons over people ). Similarly, Oenbring (2010) notes that Standard Bahamian relies heavily upon passi ve voice constructions (see, for example, table 4), something that the author reads as a technique to increase the appearance of formality. To a grammarian and/or a linguist, a passive voice sent ence is a sentence in whic h the agent, that is the person or thing doing the action, isn’t the grammatical subject of the sentence.1Accordingly, passive voice constructions can allow the write r to avoid imparting individual volition on (what is referred to by linguists as agency ) or avoid even naming the pe rson doing the action. In violent crime reporting, passive voice cons tructions can hide or downplay the role that government actors such as the police, EMTs, or magistrates play in causing an action to happen. See, for example, the following pairs of constructions, one pa ssive and one active. No te that all of these passive voice constructions appeared in the BVCC n-grams. Passive 1: The victim was taken to hospital Active 1: EMTs took the victim to hospital. Passive 2: The accusedswere not required to enter a plea. Active 2: The magistrate required the a ccuseds to enter a plea. Passive 3 : It is alleged that Mr. Rolle did something. Active 3: The police allege that Mr Rolle did something. As table 3 and 4 suggest, violen t crime reporting in all locations displays a greater rate of passive voice constructions than general newspa per reporting discourse in the same respective 1 Active voice: Ms. Rolle ate the conch salad. Passive voice: The conch salad was eaten by Ms. Rolle.


Draft: Violence Symposium 3rd November 2011 Page | 11 locations (table 4 is taken from Oenbring [2010]). What’s more, Caribbea n variants of English display much greater use of passive voice than British or American variants in both studies. In general newspaper writing and violent crime repor ting, the Bahamian corpus has the highest rate of passive voice constructions. Table 3: Passive Voice Frequenc y in the Violent Crime Corpora Area Corpus size Percent Passive Voice Bahamas 19,000 words 36% Trinidad 17,000 words 33% Jamaica 20,000 words 31% South Florida 21,000 words 19% Table 4: Passive Voice Frequency in Gen eral Newspaper Corpora (from Oenbring [2010]) Percent Passive Voice Bahamian 23% Jamaican 20% American 5% British 4% The preference for passive voice constructions in the BVCC stems, for certain, in large part from the general preference for passive voice constructions in formal Bahamian English.


Draft: Violence Symposium 3rd November 2011 Page | 12 However, one might also speculate that unwillin gness to call out or pin down government actors in regards to their actions -perh aps in the interest of politeness -also plays a role in the high rate of passive voice constructions in the BVCC. Indeed, a common critique of writing heavy in passive voice constructions is that can lead to ‘bureaucratese’ (e.g., the pa ssive voice construction I am advised ). Keywords Table 5 presents the top 20 keywords for th e BVCC with different reference corpora used as baseline. I have placed the particularly inte resting terms in bold. Note that the corpus linguistic software (in this case WordSmith) onl y produced seven keywords for the BVCC when the TVCC was used as the reference corpus. Table 5: BVCC Keywords with Different Reference Corpora Rank JVCC SFVCC TVCC 1 AROUND YESTERDAY MEN 2 OFFICERS ROAD ARMED 3 ARMED ARMED MS 4 CAR MS SKIPPINGS 5 STREET CASH STAFF 6 POLICE ON CASH 7 MS INVESTIGATIONSFREEPORT 8 ROBBERY MAGISTRATE 9 CONDITION ARE 10 STABLE MEN 11 SKIPPINGS SKIPPINGS 12 OFFICER FREEPORT 13 AT AROUND 14 ROBBED AREA 15 CASH TAKEN 16 MEN AMOUNT 17 HANDGUN MR


Draft: Violence Symposium 3rd November 2011 Page | 13 Rank JVCC SFVCC TVCC 18 SMALL BAHAMAS 19 MR STABLE 20 BAHAMAS RESIDENT Of course, all three lists of keywords have several terms re lating to the fact that BVCC takes place in the Bahamas that are not particular ly interesting from a linguistic perspective. These terms include: Bahamas, Freeport, and Skippings (the surname of a particular officer). More interestingly, Oenbring (2010) finds that Standard Bahamian English has a particular sensitivity to gender and social roles. This sens itivity to social titles and roles likely accounts for the presence of Ms. and Mr. in the BVCC keywords.Following th e content analysis of the current study, several of keywords indicate greater reporting of armed robbery in the BVCC in comparison to the reference corpora. Suggestive terms include: armed, robbery, robbed, and cash. Moreover, following the N-gram analysis of the current study, several BVCC keywords are indicative of stock expressions, including: sable, condition, and taken (to hospital or into custody). Other common terms in the BVCC include around (to indicate the time of the incident) and amount (almost always in the stock phrases a small amount of cash or an undetermined amount of cash ). Discussion Although by no means attempting to be a final word on the subject, the current study is a first tentative step toward a more comprehensive understanding of representations of violence in Bahamian media. While the author recognizes th at the two mainstream newspapers in Nassau may not play a primary role in regulating di scussions of violence in The Bahamas — and, moreover, he recognizes that other forms of media including music, television, and radio may


Draft: Violence Symposium 3rd November 2011 Page | 14 play a much more important role in framing di scussions of violence in The Bahamas — he does believe the findings of the current article to be a valuable rejoinder to what he hopes will prove to be a lively discussion regarding the effects violence in the media have on Bahamian society. As to the question of whether Bahamian newspa pers sensationalize violence, the current piece finds no significant evidence, at least on the level of language choices, that the mainstream newspapers use overly emotive or embellished langua ge to report acts of violence. Indeed, if anything, the language of mainstream Bahamian ne wspapers appears more staid and perfunctory than sensationalistic. References Biber, D., Conrad, S., &Reppen, R. (1998). Corpus Linguistics: Investigating Language Structure and Use. Cambridge: Cambridge UP. Bahamas Can’t Hide Crime Problem. (2009). The Nassau Tribune. Retrieved from t/12172009_crime_editorial_pag4 Chilton, P. (2004). Analyzing Political Discou rse: Theory and Practice. London: Routledge. Chouliaraki, L., &Fairclough, N. (1999). Discourse in Late Modernity: Rethinking Critical Discourse Analysis Edinburgh: Edinburgh UP. Fairclough, N. (1992). Discourse and Social Change. Cambridge, MA: Polity Press. Fielding, W. J., Oenbring, R. A., Brennen, S., Carro ll, M. C., Bethel, N., &Minnis, J. (2011).A first look at harm toward animals by Bahamians in childhood. The International Journal of Bahamian Studies, 17 (2), 27-49. Retrieved from Gee, J, P. (1999). An Introduction to Discourse Analysis: Theory and Method New York: Routledge.


Draft: Violence Symposium 3rd November 2011 Page | 15 Jaworski, A., &Coupland, N. (Eds.) (1999). The Discourse Reader. New York: Routledge. Krippendorff, K. (2004). Content Analysis: An Intr oduction to Its Methodology. Thousand Oaks, CA: Sage. Major, A, and Atwood, E. (2004). Environmenta l Stories Define Problems, Not Solutions. Newspaper Research Journal. 25 (3). Mills, S. (1997). Discourse New York: Routledge. Nicolls, N. (2010). How far can media go in search of truth?. Retrieved from nsight/08162010_Noelle-s-Insight_Insight Oenbring, R. A. (2010). Corpus linguistic studies of Standard Bahamian English: A comparative study of newspaper usage. The International Journal of Bahamian Studies, 16 51-62. Retrieved from http:/ / Stubbs, M. (1996). Text and Corpus Analysis. New York: Blackwell, 1996. Wodak, R., and Michael Meyer. (Eds.) (2001). Methods of Critical Discourse Analysis. London: Sage.