June 29, 1995

Specifying Event Content
in Narratives 1

David R. Heise
Department of Sociology
Indiana University
Bloomington, IN 47405


This chapter defines a frame for specifying the content of productive social events in terms of an agent, action, object, instrument, alignment, setting, product, and beneficiary-experiencer. With the aid of a computer program, the Event Frame is used to code 61 events in a lynching narrative, and a number of analyses of the narrative's content are conducted. The empirical example illustrates the mechanics of Event-Frame coding and some kinds of analytic benefits that accrue from this form of analysis.

Qualitative researchers get considerable guidance from the research literature on how to analyze interrelations between events in a narrative (e.g., Abell, 1987; Abbott, 1992; Corsaro and Heise, 1990; Doreian, 1993; Griffin, 1993; Heise, 1989; Ragin,1987; Ragin, Mayer and Drass, 1984; Tesch, 1990). However, little guidance is available on how events themselves should be rendered so that their descriptions are adequate to the analytical demands placed on them. In fact, the major approaches to narrative analysis actually draw attention away from semantic problems in describing narrative events, in favor of analyzing relations among the events (Heise, 1993). In this chapter I begin to redress the abandonment of content in formal analyses of narratives. This chapter focuses on what needs to be said about each event in a narrative in order to allow a reader to understand what happened, to disclose logical interrelations between events, and to reveal functional linkages among the various people and things involved in the narrative episode.

Journalists confront similar problems of writing adequate descriptions of events. The guideline taught to apprentice reporters is the "5 WH formula" of Who did What , When , Where , Why , and How (Charnley, 1975: 186-7). Including all this information in the first paragraph of a news article results in what is called an Associated Press lead. Consider this hypothetical illustration: "Homeowner Michael Gordon, 300 Air Lane, held a burglar at gunpoint early this morning until police arrived to make a formal arrest." Who is specified by the name and by the role of homeowner. What is the action of holding the burglar so he cannot flee. When is early this morning (the actual date being available at the top of the newspaper page). Where is given in the address and implicitly in the setting-linked role of homeowner. Why is specified by reference to linked events -- an attempted burglary and police arriving to make a formal arrest. How is understood from the reference to gunpoint.

Journalists who artfully use the 5 WH formula pack a single sentence with enough information so that a reader can envisage what happened. Such a powerful expository device deserves adaptation for social scientists who are trying to write clear and useful narratives. That is the aim in this paper. In essence, the 5WH formula will be elaborated into a framework for writing well-composed social science descriptions of events in which actors use various means to transform people or things from one condition to another (for example, events like a physician using drugs to medicate an ill patient in order to make the patient well). The elaborated framework reduces somewhat the level of artfulness that is required for writing good event descriptions. More important, it helps assure that every event description in a narrative includes elements that are involved in logical interrelations between events, and that can be used to reveal functional linkages among the various people and things involved in an episode. I use a 1930 lynching as an empirical illustration to show that formulating events with the framework can be the basis for analyses that help identify logical and functional connections.

The Event Frame

Elaboration of the 5WH formula is accomplished by departing from journalists who write sentences describing events and turning to linguists who analyze such sentences. Case grammar divides a sentence into meaningful parts of speech like subject, verb, and object -- units that are known to have cross-cultural universality (Greenberg, 1963). Charles Fillmore (1968) argued that "The case notions comprise a set of universal, presumably innate, concepts which identify certain types of judgments human beings are capable of making about the events that are going on around them, judgments about such matters as who did it, who it happened to, and what got changed" (Dirven and Radden,{2} 1987, p. 24). Over a period of two decades Fillmore and other case grammarians (e.g., Chafe, 1970) have elaborated the syntactic-semantic representation of sentences. This body of work permits formulation of a sociological Event Frame, which has useful faculties for guiding formal qualitative research.

The sociological Event Frame recapitulates the system that humans everywhere use to conceptualize events, and the Event Frame represents a generalized implementation of the ethnomethodological suggestion that studying social structure requires analyzing the ' content of talk in its semantic aspects.' Anyone, regardless of the language they speak, makes meaningful statements about productive social events by explicit or implicit use of certain categories of meaning, and sociologically elaborated, these categories provide a formal basis for describing events. (Heise and Durig, forthcoming)

Table 1 summarizes the Event Frame. The action category corresponds to the verb in an event description. The other seven categories correspond to nouns, or more generally to phrases that identify entities that explicitly or implicitly arise in the event description. Entities may be persons -- e.g., agent and beneficiary always refer to a person or an aggregate of people. Entities also may be non-human things, which divide into physical entities and semiotic entities. A physical entity is a reference to a particular object, place, or time -- e.g., the setting of an event always refers to a physical entity. The semiotic entities that arise in narratives are verbal references to past events -- e.g., a woman reporting to a newcomer, "They hit my husband" -- or to future events -- as when a person makes a request by specifying the desired action, "Give me the keys."

The case definitions in Table 1 are fairly faithful to linguistic formulations (Heise and Durig, forthcoming), so I will add some sociological perspective to them in the following notes in order to facilitate sociological applications.{3}

Table 1. The Event Frame
Agent The instigator of a happening. Action The fusing of event-frame elements into a happening.
Object The entity whose condition is expended by movement or change during the happening, such that a repetition of the happening would require replacement. People can be objects.
Instrument An entity that used by the agent causally to advance the happening while not being significantly changed by the happening. People, social organizations, and verbalizations can be instruments.
Alignment The specific place or time at which an instrument is applied to an object or in a setting.
Setting A convergence of relatable agents, objects, and instruments within a space-time boundary.
Product An entity that comes into existence as a result of a happening and that enables or disables subsequent happenings.
Beneficiary / Experiencer A beneficiary is the agent of an event that intentionally is enabled or disabled by the agent in the focal event. An experiencer is a perceiver whose knowledge or impression of things intentionally is changed by the agent of the event.

Agents generally are specified in terms of social identities or roles. This is in the interests of generality: e.g., observing that a "foreman" did something offers more opportunity for sociological generalization than saying "Bill Jones" did it. Additionally, specifying a role suggests the competencies and other personal resources that the agent has for accomplishing the given action. Collective or organizational agents can be specified, in which case all behaviors that contribute to an action are attributed to an encompassing group or corporate identity even though performed by different people.

Action has a somewhat nonintuitive meaning in the Event Frame due to its origins in linguistic analysis, where it corresponds to a verb. Naming an action only indirectly suggests the body motions and manipulations of things that Habermas (1989) called "concomitant execution" of an action, the implicit understanding being that a competent agent takes care of behavioral details. Rather than being merely descriptive in a time-motion sense, an action identifier functions as a keystone of the event, holding other elements together like a Gestalt and putting constraints on what the other elements might be (for example, a named action constrains possible agents, since the agent must have the required competence). Naming productive actions is the means by which we insinuate causality. A productive action implies possible agents, objects, instruments, alignments, and settings, and it evokes expectation of a particular product. Consider for example the productive action of "shooting" someone, which arises later in the empirical illustration. As used in contemporary America, this verb implies a gun as instrument, a human with hands beyond infant development as agent, an organism capable of agency as the object, the body of the organism as an alignment, a setting with physical dimensions corresponding to the gun's range, and an impaired organism as product. The manifestation of "shooting" is the entire event -- to say that one man intentionally shot another means that the two men were within sight of one another and the first man, by pointing a gun at the second man and squeezing the gun's trigger (concomitant execution), transformed the other man to an injured or dead person.

Objects are the focus of concomitant actions and the items that are transformed into products in productive events or expended in experiential happenings. Objects of events are used up in some sense or relocated, so the object has to be replaced before the same event can occur again. Some events operate on multiple objects, with one object being deemed focal while the others are viewed as supplies. A semiotic object is a message or text that instigates concomitant action on supplies to produce a product that manifests the semiotic message -- e.g., a cook processing a diner's order yields a restaurant meal.

Instruments are entities that are applied to objects in order to yield products. The distinction between instruments and supplies is that instruments are not used up by an event and are available without replacement for reoccurrences. Heise and Durig (forthcoming) emphasize that other people and whole social organizations can be instruments of individual action. For example, "the parent took the sick child to the doctor" means that the parent used a medical establishment controlled by a particular physician as an instrument to transform the sick child into a well child.

An alignment is a reference to the part of an object or the part of a setting where an action takes place. For example, "nose" is the alignment in "The mother kissed her child on the nose." Alignments can refer to time, as does "previews" in "The boys at the movie threw popcorn at each other during the previews." People describing an event often do not specify alignments if they are part of the familiar concomitant execution of an action -- it seems pedantic to do so (e.g., we ordinarily say "hammered the nail," not "hammered the nail on the nail's head with the hammer's head"). However, familiarity is a relative matter, and sociological descriptions of events can gain useful information by attending to alignments.

A setting is the space-time arena where agents, objects, instruments, and alignments are assembled and functional for the purposes of an action. A setting name may focus on place -- e.g., the White House, or time may be focal, as in describing events at Christmas. Settings may enclose other settings -- e.g., the Oval Room in the White House. The appropriate reference is to the smallest setting that actually contains agents, objects, instruments, and alignments for the action under consideration. Settings can be quite large with enclosed settings interpreted as alignments: e.g., the setting for "The woman drove from home to work" would be the town where she lives, or at least the section of town that includes her residence and workplace since these are specified alignments in the event. Settings sometimes originate as products of action, and they can become objects of action, as in "The interior decorators painted the Oval Room."

Products are the entities that an action effects by transmuting, reorganizing, re-locating, or re-configuring objects. Products are material things that can participate in other events -- e.g., as an object or instrument or setting. Thus the emergence of a product is an objective fact in the sense that the product enables some other event that could not occur without that product. On the other hand, cognizance of a product is culturally based, so emergence of a product also has a subjective component. For example, Holy Water produced by a priest's blessing is objectively real in that the Holy Water serves as an essential supply for later events, but the product is subjective in the sense that plain water is transmuted to Holy Water only for those people who attribute supernatural significance to the rite. The mere experience of an event ordinarily is not a product, but there is a product if the experience changes the capacities of the experiencer. For example, "the boy showed his friend how to use the sling shot" socializes the second boy into an enhanced agent with some sling-shot competence; and an emotionally traumatic experience, like being kidnapped, can impair agency and create a human who becomes an object of psychiatric treatment. An event can have multiple products, some of which may not be purposeful, like the cloth remnants produced in tailoring. In general, though, productive events are implemented in order to yield products that enable (or disable) other events through the provision of agents, objects, instruments, alignments, or settings.

Beneficiaries are people who the agent knows may use the event's product in a subsequent event. The agent creates the product for the beneficiary and takes the beneficiary's standards of usage into account in generating the product. A beneficiary in the Event Frame always is manifest. Someone who uses a product without the agent's knowledge is not a manifest beneficiary since that person does not enter into the agent's construction of the original event (such a person could be referred to as a latent beneficiary). While the connotation of the word "beneficiary" is positive, the Event-Frame usage is neutral, and people can "benefit" from products that they use under coercion (e.g., the victim would be the beneficiary of the kidnapper's action in "the kidnapper gives the victim handcuffs so that the victim can shackle himself"). The Event Frame category for beneficiary also includes experiencer , because some events like entertainments are conducted simply so that a particular person can experience the event, and the event does not yield a product that the experiencer will use. An Event-Frame experiencer is manifest in the sense of being acknowledged by the agent, and the agent takes the experiencer's standards into account in executing the action. An observer who is unknown to the agent is not a manifest experiencer; rather the unknown observer is the agent of a parallel event that has the first event as object. For example, students are experiencers in "the scientist conducted the test for the students," but the spy is not an experiencer in "the spy watched the scientist conduct the test." Any event with a manifest beneficiary or experiencer can be called "purposeful."

Determinateness of the Event Frame

The authority of the Event Frame derives from the fact that people use vocabularies of agents, actions, objects, instruments, etc. to construct sentences describing events, and languages have syntactic rules defining how these classes of meaning have to be distributed in sentences. The Event Frame simply organizes the conceptual resources that have evolved in human languages into a data structure that can be applied methodically in sociological analyses of happenings.

Yet each semantic-syntactic class is itself a human construction based on observations of speech, so the detailed definitions of the classes must be viewed as open-ended and subject to improvement. Linguists approach the problem of improving these definitions by trying to refine the semantic rules delimiting vocabulary lists for each category and the syntactic rules specifying how words in a list must be used. Durig and I took a different approach, trying to adjust the basic linguistic definitions in such a way as to be sociologically provocative and productive. For example, the definitions provided here, with an added sociological axiom, offer a logical basis for understanding how collaborative action in an organization gets so predictable and determinate that it can be used instrumentally by outside individuals (see the discussion of macroaction in Heise and Durig, forthcoming). Still another way of refining the definitions is by practice with them in applications. This already has resulted in changes that have been incorporated into the definitions given here, and the empirical example in this paper will suggest some areas for future work.

The Connections Program

The Event Frame indicates what items of information have to be recorded in order to achieve a thorough, analytically useful description of an event. A narrative in which key events are described in this detail would be less ambiguous than reports commonly are, and it would provide a good deal of data for studying the interrelationships of people and their linkages to entities other than individuals.

My first ventures in applying the Event Frame to narratives proved that the coding process is tedious. In essence, one takes each sentence from the original narrative and expands it, filling in the slots of the Event Frame by typing the agent, action, object, etc. The same items (e.g., agents) appear repeatedly in different events, so one has to type the names of these items over and over. Moreover, some items, like an unchanging setting, may be mentioned only at the beginning of a narrative, yet the Event Frame requires re-entry for each event. All of the coding does produce a rich body of data, but when the coding is finished, another problem arises. The data are qualitative, unsuitable for analysis with standard statistical computer programs, so it is difficult to make analytic use of one's work.

Retaining faith that the Event Frame could be a useful sociological device, I addressed these problems by developing a computer program to eliminate the tedium and to offer some analytic possibilities for use on completed datasets. The resulting Connections program, which runs on Macintosh computers, removes most of the repetitiousness from the coding process and allows one to use the final dataset to answer some sociological questions about how people, things, and actions were linked over the course of the narrated episode. The following procedural description outlines the basic operations involved in using Connections . I will elaborate on substantive issues involved in coding and analyses in the empirical example.

The first step is to set up a narrative text that will be the basis for detailed specifications of events. The text may be imported from a word processing program or it can be entered directly within the Connections program. Thereafter, coding of events proceeds by repeated application of the following routine.

The analyst selects a block of text in the narrative that identifies an event -- often a sentence but sometimes just a phrase, sometimes multiple sentences. That reference text automatically is copied to a box titled "Source Text" on an event-definition screen, and the analyst enters a short sentence describing the event in a box titled "Event Name." A new sentence may sound like a redundancy, but that often is not the case, since the source text merely may allude to the event, and part of the coding process is to provide a sentence description in which the agent, action, and object are explicit, with some careful thought given to selection of an action word.

At the right side of the event-definition screen are eight boxes, labeled with the names of the Event Frame categories. These boxes are used to specify the event's composition in detail. The entry in the Action box is a copy of the action word in the Event Name. Entries in the other boxes are accomplished by making a selection from one of two menus, one listing people at the scene and the other listing non-human entities. Items are added to the two menus by shifting temporarily to a person-definition screen or to an entity-definition screen and typing the names -- thereafter person names and entity names are selected from the menus. Multiple items can be included in any category. Having completed the coding of one event, the analyst returns to the original narrative, and makes another selection in order to code another event.

The current analytic capability of the program allows one to ascertain the linkages of any specified person, entity, or action to all persons, entities, or actions in the narrative. Linkages can be restricted to a particular functional type -- for example (alluding to the empirical example of a lynching narrative that follows), one can answer a question like, "Who were beneficiaries of the sheriff's actions?" or "What were the products of the mob's actions?" Moreover, one can code abstraction hierarchies of people or entities and use this information to aggregate specific linkages into more general relations. Say, for example, that the business man, lawyer, and concerned Whites in the lynching narrative are coded as "the accused man's supporters"; then one could answer a question like, "What were the instruments used by the accused man's supporters?" and instruments used by the business man, lawyer, and concerned Whites all contribute to the answer.

The Connections program makes Event-Frame coding easy and rapid, and it provides some rudimentary analytic capabilities. I forgo further discussion of the program in this exposition. What follows is an empirical example that reveals some of the problems and benefits that come from coding narratives with the Event Frame.

An Empirical Illustration

The Event Frame can be applied to any kind of narrative, but the strength of the scheme is most evident in dealing with narratives about purposeful productive events because then rich data accumulate in all eight coding categories. This is in contrast to analyses of a conversation or discussion where variations occur in agents, objects, and experiencers, but other elements in the Event Frame are absent, constant, or difficult to define. (That is, a face-to-face conversation may have no perceivable product, or it may result in person changes that are difficult to articulate; the actions are limited -- "says," "asks," "replies," etc.; setting is constant; alignments change little; and the instrument either is a constant -- "speech" -- or else participants' talk must be dissected in terms of prosodic and other conversational devices.)

For an empirical illustration, I chose to follow Griffin's (1993) lead in focusing sociological attention on lynchings during the Jim Crow era (1930). Any gain in understanding of lynchings might have important ramifications in dealing with racism in contemporary America. Moreover, lynchings are short, dynamic episodes packed with social processes so they are appropriate for Event Frame analysis. An analysis of past events inevitably is seen through multiple layers of subjectivity -- in this case, my formulation of an author's rendition of written records of what people said had happened. The Appendix to Griffin's (1993) paper provides a thoughtful discussion of the fallibilities of using this and other historical records as a basis for sociological analyses.

A source used by Griffin, Arthur Raper's The Tragedy of Lynching (1933), provides manageable accounts of numerous incidents. I selected a case that began on Easter Sunday, 1930, in Walhalla, South Carolina. On that day, a White woman accused Allen Green, a prosperous 52 year old African-American male, of seizing her and manhandling her from porch to kitchen, binding her with belts to a chair, and raping her, while her husband went to the front of house to fetch his waiting sister. This alleged incident occurred during a 9:30 a.m. business visit to Green's house. Green was arrested, given a preliminary hearing in court, and held in the sheriff's home which served as the county jail. Mobs assembled on Monday, Tuesday, and Wednesday nights. At least one prominent White tried to pacify the crowd, but the Mayor of the town -- a long-time enemy of Green's -- agitated and encouraged violence. On Wednesday night the mob stormed the sheriff's house, clubbed the sheriff, seized Green, drove him to the countryside, and shot him to death.

I copied Raper's (1933) two-page narrative of events that followed the alleged rape into the Connections program, inserting some of Raper's later comments parenthetically when they helped identify the sequence of happenings. Then I coded segments of text from the narrative into brief descriptions of 61 events, along with specifications of how each event was constituted in terms of the categories within the Event Frame. The resulting dataset is too long to present in its entirety, but some sample codings are presented in Table 2.

Coding Considerations

Though I am focusing on the lynching, I included the alleged rape in the analysis because the event is referenced by people in the narrative, and through their rhetorical activities the event became part of the constituted reality that led to the lynching. Additionally, I include it in Table 2 (Event 01) because this event offers the opportunity to discuss some aspects of Event Frame coding.

Table 2. Event-Frame Codings for Some Sample Events

Event Agent Action Object Instrument Alignment Setting Product Beneficiary
01. Green rapes white woman, allegedly. Green rapes White wife razor, belt Green's house Green,
05. Husband's father tells sheriff of incident. vengeful White man tells sheriff 'Green rapes' jail (sheriff's home), day of alleged rape investigator sheriff,
06. Sheriff takes self to victim's home investigator takes investigator jail (sheriff's home), couple's home town of Walhalla , day of alleged rape RELOCATED OBJECT investigator,
10. Sheriff and deputy arrest Green. policeman with arrest powers arrest Green policeman with subjugation power Green's house town of Walhalla, day of alleged rape prisoner judge,
19. Mill cops warns sheriff of lynching. mill policeman, mill watchman warns sheriff 'Green endangered' jail (sheriff's home) town of Walhalla DEFECTIVE PRODUCT sheriff,
27. Wednesday, vengeful white men and boys assemble at park by mill. vengeful White man, town's night policeman, boy from nearby town, man from nearby town, filling station operator, policeman in nearby town, adolescent boy, mayor, straggler, hill billy assemble vengeful White man, town's night policeman, boy from nearby town, man from nearby town, filling station operator, policeman in nearby town, adolescent boy, mayor, straggler, hill billy automotive vehicle park by mill Oconee County, three days after alleged rape angry White mob lynch instigator,
59. Drivers light up Green with headlights. vehicle driver light up prisoner automotive vehicle prisoner lynch site, three days after alleged rape visible target, concealed executioner executioner,

I coded the rape as a non-productive event, in the sense that the perpetrator reputedly was not trying to produce a product that would enable any later event, but rather was supposed to have engaged in the action merely to experience the event. A rape could have a product like a diseased or pregnant woman, but that leads to a different story than the one under consideration (the Raper book does include some narratives of that type). The woman's rape story in this case offered no indication that Green was trying to accomplish anything beyond the experience of the event itself. Thus this event is coded as having no product and Green is identified as the intended experiencer of his action.

In principle, the rape could have been divided into a sub-sequence -- according to the woman, Green took the woman from the porch to his kitchen under threat of a razor, gagged her and bound her into a straight chair with three belts waiting on the kitchen table, had sexual intercourse with her, then released her and gave her some trinkets, whereupon the two of them returned to the porch and conversed with the husband who had watched the intercourse through the kitchen window. (The husband claimed that the fettering sequence occurred in a minute and half while he went to the front of house to fetch his waiting sister.) I treated a whole paragraph of text as a single event, merely alluding to some of the details by specifying instruments of razor and belts, because I chose to analyze the lynching rather than the alleged rape.

Alignments are not specified for the rape. Raper's text treats the alignments of a rape as well-known aspects of concomitant execution that need not be mentioned. At the same time his text (1933, p. 265) undermines the credibility of the rape story by insinuating the required alignments were impossible: "One [belt] he placed around her body over her arms and around the back of the chair; the other he strapped tightly across her lap over her dress and under the chair seat. Then he assaulted her as she sat there tightly bound." The truth value of this description seems to be undermined because we are unable to imagine the accomplishment of the concomitant execution. Evidently the sheriff was similarly perplexed because he is quoted (p. 267) as saying, "I couldn't believe anybody really thought Green guilty." (Perhaps doubt did not arise for others in the community because they never heard the woman's full description of the incident.) In any case, this example suggests an intriguing linkage of alignments to the perceived truthfulness or falsity of an event description -- that is, a description that precludes required alignments for an action seems like a fabrication to knowledgeable appraisers. The matter warrants future inquiry.

The next event (05) included in Table 2 describes the woman's husband's father reporting the incident to the sheriff. His hearing about the rape in a prior event turned him into the "vengeful White man" who is the agent in this event. (I used the modifier "vengeful" for anyone who acted against Green, as a result of hearing about the rape.) Event 05 effects another such transformation, a report of a crime transforming the sheriff into an investigator who will go and query the woman about the alleged rape. My analysis of the lynching suggests that the human products of events often are people in particular roles and motivational states, ready to instigate subsequent events. I experimented with different methods for representing motivational transformations -- e.g., "vengeful White man" as contrasted with "investigator." It is sociologically satisfying to say that the product of Event 05 is sudden salience of the sheriff's subrole of investigator. Yet the new identity makes it seem like a new person has entered the narrative when we just have the same person in a new role. The other convention of using modifiers -- a vengeful White man, an investigative sheriff -- may be the better approach. At any rate, the question of how to represent a person's multiple identities in Event Frame analysis remains an unresolved issue{4}.

In Event 05, the instrument is a semiotic entity consisting of a verbal reference to Event 01, as symbolized by a short name for Event 01 in single quotes, 'Green rapes'. That is, the vengeful White man instrumentally applies the story of criminal rape to the sheriff in order to turn the sheriff into an investigator -- a role needed for subsequent events by the sheriff. (A more detailed analysis perhaps could show that this instrumental story also is providing alignments for the sheriff's future actions, leading him to the home of the victim and the home of the perpetrator.)

I include Event 06 in Table 2 to show how I handled the frequent case of locomotion from one place to another. The agent operates on the self as object (and often with the self as instrument as well, using the feet for walking), effecting a relocated self as product so that the self as beneficiary can engage in further events. The beginning and ending positions are specified as alignments, and the setting is the culturally recognized spatial unit that includes these positions. I used the word "takes" for the action, though "moves" would be an alternative. Other kinds of motion acts also arise in narratives -- e.g., an agent moves her hand from x to y, or an agent drives another person from place to place.

Event 06 takes the sheriff to the White woman's home, and the next event reported in Raper's narrative is the woman telling the sheriff about the alleged rape. The sheriff presumably queried her to elicit this response, but the query is not in Raper's text or in my analysis. The question of how many presumed events should be added to the explicit narrative is an unresolved issue. Textual presentations skip over anything that readers can figure out for themselves, but leaving such events out may bias analyses based on Event Frame codings. For example, if I had included the sheriff querying the woman, analyses of women's functions in this narrative would show one more instance of a woman being an object rather than an agent in events.

Event 10 in Table 2 shows the sheriff (now a policeman with arrest powers) arresting Green through the deployment of his deputy (a policeman with subjugation powers). This is an example of an agent using another person as an instrument of action. "Arresting" is a verb with potent semantic implications: it requires a law officer as agent, and the action converts the object person into a prisoner so that a judge can act on that person. Thus the product and beneficiary can be coded even though they are not mentioned explicitly in the text.

The agents in Event 19 in Table 2 are two law officers who worked near the park where the mob was assembling. The action -- warns -- is entered ungrammatically to keep the coding comparable to other instances of warning where the agent is singular rather than plural; analyses in the Connections program would treat "warns" and "warn" as two different actions. The two agents in Event 19 are trying to use a semiotic instrument (their report on the mob) to bring out the sheriff's protective subrole so that he will take Green away. They fail to achieve their purpose, so I coded the product as "defective."

Event 27 in Table 2 shows a variety of people mentioned explicitly in Raper's text assembling themselves on the night of the lynching and thereby creating a collective agent, an angry White mob, probably purposely so the mob could be agitated by a lynch instigator. In principle, the long list of individual agents could be used wherever "mob" appears in coding, but use of "mob" is true to Raper's textual description, and the usage allows glossing over innumerable milling and conversational events that kept the individual agents together and collaterally motivated. Actually, it is impossible to avoid glossing these details of crowd behavior since they are not reported in Raper's narrative (or, for that matter, in most other descriptions of collective action). Yet, in glossing those events through use of a collective agent, we inevitably are understating the extent to which the mob members moved about and conversed with one another.

Event 59 in Table 2 refers to drivers of trucks and cars arranging their vehicles so as to light up Green, who was tied to a tree. I coded this event as having two products. The first product was a visible target for subsequent shooting. If this interpretation of the production is correct, then it was at that moment that Green's captors depersonalized him from a human prisoner to a mere material object. That allowed the multiple executioners in the next event to understand their shooting as a kind of target practice rather than murder -- "a single shot was fired, then a volley, perhaps a hundred shots" (Raper, 1933, p. 267). I also made use of an "insight" and said that the intense spotlighting of Green served to create concealed executioners. This accords with some evidence that the lynchers desired anonymity -- "They [the mob members] wanted one hundred, some seeming to have the idea that if as many as a hundred participated the lynching would be legal" (Raper, 1933, p. 266). However, I have reservations about including this second product. First, including "concealed executioners" as a product of Event 59 does not follow the basic rule that products are transformations of objects -- the prisoner was not converted to a concealed executioner. A better way to handle the insight would be to add another event in which the vehicle drivers directed their headlights so as to cast potential executioners into the shadows. That, however, punctuates my second reservation -- that hiding executioners was not mentioned in Raper's text as a recorded intention of the vehicle drivers. The general unresolved issue is: Should one add events beyond the explicit narrative to incorporate insights into what actors "must" have been thinking? Perhaps the answer is "Only when such events are necessary to enable subsequent recorded events," but this matter deserves continuing debate in discussions of historical method.

Logical Linking of Events

Event structure analysis (Heise, 1989; Corsaro and Heise, 1990; Griffin, 1993) yields generative models of event sequences by having an analyst decide what events are prerequisites for others. An analyst judges whether one event implies another on the basis of knowledge about other cases and on the basis of various kinds of theory. Additionally, many decisions about prerequisites seem to be straightforward inferences from information that exists within the narrative alone. Application of the Event Frame reveals why some decisions about implication can be answered from the narrative alone.

Consider the two events: 06 Sheriff takes self to victim's home, and 07 White wife tells sheriff of incident. In an event structure analysis, the analyst has to answer whether the first event (or an equivalent happening) is required for the second event. The answer in this case is a straightforward yes, because the woman cannot tell the sheriff her story while she is at her home and he is at his home. The action of "tell" does not work at a distance, and a locomotion that relocates one of the separated parties is essential in order to create a setting where telling can be done. (Telephoning might be an equivalent happening to enable telling.) The linkage between these two events could be made more explicit by saying that the first event yields an interview situation as its product, and the second event occurs within the setting of that interview situation.

Another illustration of straightforward linkage involves the events: 26 Crowd arms itself, and 60 Executioners shoot Green. The first event is required for the second event because the first produces armed men, and men with access to firearms as instruments are essential for the second event to occur. In this instance, the product of the first event provides the instrument for the second event.

Another variation is provided by the two events: 07 White wife tells sheriff of incident, and 10 Sheriff and deputy arrest Green. In the first event the woman claims that a criminal act occurred, thereby evoking the sheriff's subrole as an agent of the court with arrest powers. Then the sheriff in that role is the agent who arrests Green in the subsequent event. Some event like 07 had to confront the sheriff with decisive evidence of a criminal action in order to induce an appropriate agent for Event 10. The American legal system allows arrest powers to be mobilized only for just cause.

A fourth example involves the two events: 59 Drivers light up Green with headlights, and 60 Executioners shoot Green. The first event created a visible target in the darkness of night. That visible target then became an object and point of alignment for the mass shooting in the second event. In the circumstances some event like the first was required before the mass aiming and firing of guns in the second event could take place.

These examples reveal that straightforward logical linkages between events arise when the product of a first event constitutes an element that appears in a second event, and that element would be unlikely to exist without occurrence of the first event. The prior event is a prerequisite for the later event because it provides the setting, the instrument, the agent, the object, or the alignment of the later event. The qualification is important, though. A seeming prerequisite can be invalidated because its product is so readily available that an event producing that entity is redundant and unnecessary for later events using the entity. For example, men arming themselves would be unnecessary for a mass shooting if men in the area routinely were armed with guns in their vehicles, and deciding whether personal arming is a prerequisite for the shooting requires weighing evidence about whether the lynchers would have guns available to them anyway. Such redundancy arises when some relevant productive events are outside the narrative, and were those events included in the narrative, then prerequisites could be specified disjunctively. For example, guns in all vehicles must mean that each owner armed his vehicle in the past, before the time of the narrative; and recognizing this, the shooting could be specified as having disjunctive prerequisites of vehicle arming or personal arming.

A prior event may or may not be a prerequisite if it appears to contribute no element to the later event. One problem in making negative decisions about prerequisites is that an indirect dependence of one event on another can be obscured in a long sequence of intervening events. Both Heise (1989) and Warfield (1990) applied the term "local logic" to inferences made without needing to consider intervening states, and Warfield also defined "intermediate logic" as having to allow for one or two intervening states, and "deep logic" as having to allow for three or more intervening states. As Warfield (1990, p. 93) noted, "Local logic is easy to generate in the mind, if the needed information can be found. Intermediate logic is more difficult, but certainly can be produced in the mind. Deep logic is frequently very difficult to produce in the mind, simply because the mind cannot manage to recall and work with simultaneously the necessary volume of information." Heise (1989) noted that difficulties presented by intervening events can be mitigated by breaking an inference problem down into a number of decisions involving adjacent events where judgments of dependency are easier to make. That resolution of the problem is an option offered in the computer program for conducting Event Structure analyses.

Another problem is that some materially nonproductive events might be mentally productive in producing outlooks that establish a basis for later events. To illustrate, consider the following two events that occur in the lynching sequence right after a jail attacker has struck the sheriff a disabling blow on the head: 45 Jail attacker commands 'give keys', and 46 Sheriff's wife gives keys to jail attacker. Aside from producing a frightened woman, the command establishes the jail attacker as a beneficiary for her giving away the keys -- an act she would not do purposelessly. Moreover, the command implicitly instigates another realization -- that performing the second event can disable further acts of violence by allowing the jail attacker to get on with what he wants to do. These internal psychological realizations might have been essential for the sheriff's wife to give away the keys. Thus, some events with no products or seemingly unrelated products still might be required for a focal event because they generate a particular understanding of the situation in the agent of the focal event.

Functional Relations

Event Frame coding of the multiple events in a narrative yields a database of people and things involved in the overall episode, with each item in the database categorized according to how it functioned in particular events. The database allows a number of different kinds of analyses to be done.

Consider, for example, automotive vehicles (my designator for cars or trucks). Automotive vehicles took on three different functions in the lynching narrative. In nine events they were instruments of action, where the actions were assembling, chasing, dispersing, driving, lighting up with headlights, and stopping at the lynch site. An automotive vehicle also served as a setting for two events, one in which the African-American begged and the other in which he was terrorized. Additionally, automotive vehicles were alignments in three events -- in one case vehicles were the place into which mob members crowded, in another case they were the place to which mob members dispersed, and in the third case they were the place to which the African-American was taken. Thus vehicles were crucial in this episode of lynching, enabling a variety of actions and serving as environments and points of reference in environments, beyond serving as instruments of locomotion.

Table 3

Associations of Automotive Vehicle as Instrument

Agents Actions Objects Alignments Settings Products Beneficiaries
/ 2 adolescent boy, 2 boy from nearby town, 1 concerned White, 1 filling station operator, 2 hill billy, 1 lawyer, 1 man from nearby town, 2 mayor, 2 mobilized White mob, 2 vengeful White man, 1 policeman in nearby town, 2 straggler, 2 town's night policeman, 2 vehicle driver 2 assemble, 1 chases, 1 disperses, 3 drives, 1 light up, 1 stops / 2 adolescent boy, 2 boy from nearby town, 1 filling station operator, 1 Green, 2 hill billy, 1 lawyer, 1 man from nearby town, 2 mayor, 4 mobilized White mob, 2 vengeful White man, 1 policeman in nearby town, 2 prisoner, 2 straggler, 2 town's night policeman 1 countryside, 2 home, 3 jail (sheriff's home), 3 lynch site, 1 neighboring towns, 2 park by mill, 1 town of Walhalla / 1 prisoner 1 lynch site, 6 Oconee County, 7 three days after alleged rape, 2 town of Walhalla, 2 two days after alleged rape / 4 RELOCATED OBJECT, 1 visible target / 3 angry White mob, 1 concealed executioner, 1 dispersing mob / 1 armed White man, 1 boy from nearby town, 1 concerned White, 3 executioner, 1 lawyer, 2 lynch instigator

Focusing just on the events in which automotive vehicles served as instruments, we also can address a number of questions about who utilized vehicles and under what conditions they did so. Table 3 presents the relevant results from analyses. The number preceding an entry in Table 3 registers the number of events in which that element appeared, along with vehicle as instrument. Non-human entities are listed first, then people after a slash.

The analyses indicate that a wide variety of agents employed vehicles. However, the results in Table 3 overstate the case because of the two events involving the action "assemble." All individuals who were mentioned as mob members were coded as agents assembling into a mob, but the narrative noted only some of them as using vehicles. As is often the case with an aggregate entity, something that is characteristic of the aggregate gets attributed to every member. If we restate the contribution from those two events as "some members of the mob," then the list of agents using vehicles is: 1 boy from nearby town, 1 concerned White, 1 lawyer, 1 mobilized White mob, 2 vehicle driver, and 2 some members of the mob. "Boy from nearby town," "mobilized White mob" and "vehicle driver" also can be absorbed into "some members of the mob." Doing so reveals that vehicles mainly were an instrument of mob action, and they were an instrument of those opposed to lynching only to a small degree, at least as Raper (1933) tells the story.

The list of objects suffers from the same problem as was just noted for agents. A simpler list consists of: some members of the mob, the lawyer, and Green the prisoner. Mostly these objects are being transported by vehicles, but vehicles also were used to light up the prisoner as a target.

The list of alignments shows places that people were driving to or from. "Prisoner" as an alignment refers to the event where headlights were directed at Green.

Three places appear as settings for vehicle usage: the lynch site where Green was lit up, the county from which mob members assembled, and the town where the mob moved from park to jail. Two temporal settings also appear: two days after the alleged rape, and three days after -- the day of the lynching. The temporal settings suggest that vehicles became more important instruments later in the process. Of course, vehicles may have been used as instruments on the day of the alleged rape or the day after, but they were not explicitly mentioned or strongly insinuated in any events so I did not code them as instruments.

The Products column of Table 3 shows that vehicles served to relocate objects, not surprisingly. More interesting is the fact that vehicles were the means of producing an angry White mob and of dispersing it after the lynching was completed -- males from neighboring towns drove to Walhalla to participate. Moreover, vehicles with their bright headlights were instrumental in creating the concealed executioners and visible target that were the agents and object in the actual shooting. The Beneficiaries column shows who was empowered by these products -- the lynch instigator and executioners, armed White men making their escape, boys seeking excitement. In only two recorded events did vehicles empower anyone opposed to lynching, and in both cases the beneficiary was a lawyer who rushed to the jail and then, as a concerned White, chased the mob.

This analysis indicates that cars and trucks were crucial in the unfolding of this 1930 lynching. Lynchers used their vehicles in a variety of ways that contributed to vigilante murder. Cars and trucks were an instrument of violence in 1930 (as they are today, though we disregard the fact because we are so dependent on vehicles). While other modes of transportation served some of the same functions during the peak lynching period of the late 19th century, increasing availability of automotive vehicles may have been a factor in the resurgence of lynching that prompted Raper's (1933) report.

The Connections program allows entities to be linked to more abstract categories, and then the hierarchies of abstraction can be employed in analyses to develop generalizations. Table 4 shows results when males contributing to mob action are coded as "Green's antagonists" and we ask for a profile of the events where these people acted as agents. The "Expands to" cell of the table shows all the characters who were coded as Green's antagonists, either directly or through a chain of abstractions. This list contains just characters who were unequivocally antagonistic; e.g., "armed White man" often referred to a Green antagonist, but the sheriff -- who was not antagonistic -- also appeared in the narrative as armed, so "armed White man" was not coded as a kind of "Green antagonist." Other cells of the table collapse specific characters into general types -- for example, "law officer" appears in place of any reference to deputy, investigator, mill watchman, mill policeman, policeman with arrest powers, policeman in nearby town, policeman with subjugation power, protective policeman, sheriff, or town's night policeman.

Table 4

Associations of Green Antagonist as Agent

Note. Green antagonist expands to: concealed executioner, executioner, filling station operator, hill billy, lynch instigator, mayor, mob, vengeful White man, straggler, textile worker, angry White mob, dispersing mob, mobilized White mob, relaxed White mob, identified lyncher, jail attacker, man from nearby town, White husband, White husband's father, YMCA secretary

Actions Objects Instruments Alignments Settings Products Beneficiaries
1 alarms, 1 alerts, 1 arms, 3 assemble, 1 challenges, 1 clubs, 2 commands, 2 crowds, 1 disperses, 1 holds, 1 maintains, 1 musters, 1 okays, 1 requires, 1 shoots, 1 stops, 5 takes, 2 talks to, 2 tells, 1 ties 1 visible target / 5 adolescent boy, 2 Black man, 2 DISABLED AGENT, 29 Green antagonist, 7 law officer, 1 minister, 1 people of Walhalla, 2 prisoner, 3 White female 1 'assembly', 1 'cops passive', 1 'give keys', 1 'Green rapes', 4 'mob kills', 1 'muster more', 1 'unlock Green', 1 fire whistle, 2 gun, 2 jail keys, 3 automotive vehicle, 1 rope / 1 church, 1 head, 2 home, 3 jail (sheriff's home), 2 lynch site, 1 neighboring towns, 5 park by mill, 3 automotive vehicle, 1 tree, 1 visible target / 1 fire chief, 4 Green antagonist 6 jail (sheriff's home), 2 lynch site, 5 Oconee County, 7 park by mill, 10 town of Walhalla, 2 day of alleged rape, 3 one day after alleged rape, 5 two days after alleged rape, 20 three days after alleged rape / 1 corpse, 1 DEFECTIVE PRODUCT, 6 RELOCATED OBJECT / 1 armed White man, 2 DISABLED AGENT, 12 Green antagonist, 1 Green supporter, 1 law officer, 2 prisoner, 1 White female / 1 armed White man, 23 Green antagonist, 2 law officer, 1 people of Walhalla, 2 vehicle driver, 3 White female

Table 5

Associations of Green Supporter as Agent

Note. Green supporter expands to: business man, concerned White, lawyer.

Actions Objects Instruments Alignments Settings Products Beneficiaries
1 awakens, 1 chases, 1 drives, 1 entertains, 1 identifies, 1 pleads with, 3 takes, 1 warns / 3 Green antagonist, 4 Green supporter, 3 law officer 1 'dramatic skit', 1 'Green endangered', 1 'mob disperses', 1 license plate, 2 automotive vehicle / 1 countryside, 2 home, 4 jail (sheriff's home), 2 park by mill / 1 countryside, 2 jail (sheriff's home), 1 Oconee County, 2 park by mill, 4 town of Walhalla, 1 one day after alleged rape, 7 three days after alleged rape / 1 DEFECTIVE PRODUCT, 5 RELOCATED OBJECT / 3 Green antagonist, 1 law officer / 1 Green antagonist, 6 Green supporter, 1 judge, 2 law officer

Green's antagonists came largely from the town's poorer whites who "bitterly resented" Green (Raper, 1933, p. 270), but Green also had some supporters -- "the well-to-do segment of the community -- the business and professional people -- spoke well of Green" (Raper, 1933, p. 270). Events created by Green's supporters provide an interesting contrast to those by Green's antagonists, so the events of Green's supporters are profiled in Table 5. Just a glance at the Actions columns of Tables 4 and 5 shows that Raper's (1933) narrative reports far more activities on the part of Green's antagonists than on the part of his supporters, either because Raper chose to focus on the lynchers or because the antagonists really were more active than the supporters.

The Objects column of Table 4 reveals that the antagonists mainly were operating on themselves (and this would be even more pronounced had I included the adolescent members of the mob as Green antagonists). Their next most frequent objects of action were law officers. The antagonists also directed substantial actions at the powerless -- the African-American (including his status as a target), disabled agents, and women. On the other hand, supporters attended almost equally to themselves, antagonists, and law officers.

Both the antagonists and the supporters relied heavily on semiotic messages to achieve their purposes, as revealed by the quoted entries in the Instruments columns of Tables 4 and 5. Speech was the instrument of action in about half of the events by both antagonists and supporters. Antagonists also pursued their purposes with noise (the fire whistle), guns, jail keys, vehicles, and rope. Aside from speech, vehicles and license tags were the only instruments used by supporters.

The Alignments column of Table 4 reveals Green's antagonists as busily moving between a variety of places and people and focusing their attention at sundry points. The alignments of Green's supporters in Table 5 indicates that they focused their attention less broadly. The Settings columns of the tables indicates that both groups operated within essentially the same spatial boundaries, but the antagonists were active every day from the report of the rape to the lynching, whereas supporters got active mainly on the day of the lynching.

The Products columns of the tables shows that the antagonists mainly were producing changes in their own condition or location. They also were effecting changes in supporters, law officers, females, and their prisoner. Notably, the antagonists twice disabled a person, and they killed another human thereby producing a corpse. The products of supporters' actions were repositionings, changes in the antagonists, and one change in a law officer. A person in each group engaged in some action that produced a defective product.

The Beneficiaries columns show each group mainly empowering itself through its actions. Aside from that, antagonists activated law officers, women, and townspeople. The supporters provided action opportunities for representatives of the law and once for antagonists.

Overall, comparison of Tables 4 and 5 suggests that Green's many antagonists were committed to and eagerly involved in the activities that led to Green's death. By contrast, Green's few supporters seemed to have acted reluctantly, and in any case their actions amounted to too little too late to save Green's life.

Law officers had multiple roles in the lynching. Some were on the side of law and order, arresting Green and making efforts to thwart vigilantism (though the latter were not forceful enough to be successful). Other law officers were part of the mob. I computed a profile of events with law officers as agents , and the list of beneficiaries highlights the mixed contribution of law officers. In six events they empowered agents of the legal system, but in three events they empowered Green's antagonists. Another interesting fact revealed by the analysis is that law officers used a semiotic instrument only once -- when cotton-mill police told the sheriff that Green was in danger, trying unsuccessfully to get the sheriff into a protective role. Otherwise law officers depended on guns, vehicles, and other officers to achieve their purposes.
Expands to: law officer, deputy, mill policeman, mill watchman, policeman in nearby town, policeman with arrest powers, policeman with subjugation power, protective policeman, sheriff, town's night policeman, investigator
Actions Objects Instruments Alignments Settings Products Beneficiaries
1 arms, 1 arrest, 2 assemble, 1 enlists, 2 takes, 1 warns, 1 withdraws from / 3 adolescent boy, 1 Black man, 11 Green antagonist, 7 law officer, 1 prisoner 1 'Green endangered', 1 gun, 1 automotive vehicle / 2 law officer 3 home, 2 jail (sheriff's home), 2 park by mill / 1 jail (sheriff's home), 1 Oconee County, 7 town of Walhalla, 2 day of alleged rape, 1 two days after alleged rape, 3 three days after alleged rape / 1 DEFECTIVE PRODUCT, 1 RELOCATED OBJECT / 3 Green antagonist, 2 law officer, 2 prisoner / 3 Green antagonist, 1 judge, 5 law officer

White females entered the narrative at several points. I also computed a profile of events with them as agents. In three events their actions empowered themselves, in four events they empowered law officers or a Green supporter, and in three events they empowered Green's antagonists. The females used only semiotic instruments to achieve their purposes (aside from the implicit use of their feet and hands). All of their actions occurred in their own homes or on the streets of Walhalla. I also computed a profile of events in which White females were objects or instruments. In these roles, the narrative presents them as being raped, moved, commanded, or told something. They are acted on by an African-American man (allegedly), by Green's antagonists, and by other women. The White women in the story are used to empower women mainly, but also an African-American man and a Green antagonist.
WHITE FEMALE as Object or Instrument
Expands to:, White female, frightened White woman, vengeful White female, sheriff's wife, White husband's mother, White husband's sister, White wife
Agents Actions Alignments Settings Products Beneficiaries
/ 1 Black man, 2 Green antagonist, 3 White female 1 commands, 1 rapes, 1 takes, 1 tells 1 home / 2 home, 1 jail (sheriff's home), 1 town of Walhalla, 2 day of alleged rape, 1 three days after alleged rape / 1 RELOCATED OBJECT / 2 White female / 1 Black man, 1 Green antagonist, 4 White female


Coding with the Event Frame is a way to emphasize the meanings that contribute to a narrative, thereby overcoming the retreat from substance that occurs with formal analysis of event interdependencies (Heise, 1993). Event Frame analyses highlight substantive categories and at the same time summarize how specific entities and people functioned and how different entities and people were connected. Moreover, the overwhelming detail of substantive analyses can be reduced, and generalizations obtained, through the use of semantic hierarchies.

Event Frame coding and analysis is a time-consuming process, even with the help of the Connections program. How can such laborious work be worth the effort when you get out nothing more than what is in the original narrative? The answer to this question is twofold. First, Event Frame coding demands that the analyst reorganize materials in the narrative so that they can be examined systematically and with an eye toward their general significance. Moreover, the analyst has to make explicit unwritten understandings that allow a reader to understand the story, so an Event Frame database actually reflects not what is in the narrative alone but what is in a competent, theoretically-informed reading of the narrative. Second, a moderately long narrative has far too many meanings to handle without help. Consider my reading of the lynching narrative: it included 61 events, 38 different actions, 56 kinds of human agents, 14 verbal messages and 30 kinds of physical entities. Systematic records and computer assistance are required to get command of such diverse detail and to extract inductions, like the one above that automotive vehicles were a crucial, multi-purpose instrument in the lynching of Allen Green. The payoff is illustrated by the observations and conclusions I was able to develop regarding the Walhalla lynching (which are provisional since my study is intended only as a methodological illustration).

One can improve the thoroughness of a prerequisite analysis by asking what events were required to constitute a focal event's agent, object, instrument, alignments, and setting. Tracing backwards from these elements to the events that constituted them helps in identifying prerequisite events that provide a fairly satisfying account of how the focal event occurred. Also, Griffin's (1993) strategy of considering counterfactuals when making decisions about prerequisites can be augmented by using the Event Frame. Suppose that the prior event did not occur; can we believe that an appropriate agent, object, instrument, alignment, and setting for the subsequent event still would exist? If one of these elements is lost through the counterfactual supposition, then the prior event is a prerequisite. Consider, for example, how Raper (1933, pp. 267-268) counterfactualizes activity surrounding the attack on the jail.

When the young business man reached the jail on Wednesday night just ahead of the mob, he found the front door open -- according to local citizens it is never locked -- and the sheriff asleep. When the latter came out into the hall, he found himself in the midst of the mob. Had he been awake and prepared for the invasion, it is not improbable that single-handed he could have stood them off. ... With bars between himself and the Walhalla mob, the sheriff almost certainly would have been able to make an effective defense.

In terms of the events in my rendition, Raper is suggesting that event 41 -- Mob crowds itself into jailhouse -- might not have occurred, and had it not occurred, members of the mob never would have turned into jail attackers capable of disabling the sheriff and terrorizing the sheriff's wife into handing over the cell keys. In other words, the assault on the Sheriff and his wife (and the consequent removal of Green from legal custody) depended on the mob crowding into the jailhouse because that event produced the requisite agents for the later events.

The completed database also allows focusing on specific entities or kinds of people for intensive analysis. My analysis of automotive vehicles led to the realization that cars and trucks were essential elements in the Walhalla lynching and served more functions for the lynchers than mere transportation. A comparison of events instigated by Green's antagonists and by his supporters suggested that Green died in part because his antagonists were much busier than his supporters and seemingly more committed to sustaining their activities over a period of days. Additionally, an analysis of events instigated by law officers emphasized that while some lawmen worked in the interests of order, others were empowering the agents of vigilantism. Analysis of events involving women revealed them to be limited in the settings and instruments available to them, and diverse in their commitments so that they enabled actions for every contending group.

Event Frame analysis still is maturing. For example, analysis of the lynching narrative led me to see that verbal statements can serve as semiotic instruments -- an idea that is not part of earlier statements of the framework. Additionally, as a result of this analysis I am more appreciative of the fact that some systematic method must be found for keeping track of a person as he or she moves into various roles and subroles -- e.g., the sheriff as investigator, arrester, subjugator, and protector; and a solution to that problem is yet to be found and incorporated into the Connections program. Yet even in its infancy as a sociological device, the Event Frame approach provides a useful tool for qualitative researchers.


URL: php.indiana.edu/~heise/EventContent.html
© 1997 David Heise