Janice M. Morse
A concept … refers to the properties of a phenomenon, not the phenomenon itself. It gives meaning to what can be seen, heard, tasted, smelled and touched. Thus, a concept enables us to categorize, interpret, and structure the phenomenon.
—Fawcett & Downs (1992, p. 19)
Concepts do not always emerge directly from data. Often a phenomenon emerges in the data, and it is richly described in an article but not developed (at least initially) to the level of a concept. But as these observations and descriptions accrue, eventually these studies will provide data for the development of a concept. Often these descriptions emerge slowly in articles in which they are at first tentatively addressed, perhaps not even as the main focus of inquiry. In this chapter the first example of the phenomenon of depersonalization during agonizing pain illustrates such a data source. The second example, the emergence of a concept called “preserving self,” suggested by diverse sources and disciplines, is the same process, but further along the continuum than our pain example. In both cases these data may be subsequently used to support the development of an emerging concept.
A major strategy for using qualitative data is the use of data to support the development of a concept. This may occur in two situations: The first is when you find interesting data in an interview and seek to “attach” or label it to an established concept. Here we discuss linking data with disembodiment. The second situation may occur when your research program is developing an immature or partially mature concept, and we discuss the case of preserving self.
THE DATA–CONCEPT LINK
Induction is the sacred cow of qualitative inquiry. We teach students that cognitive leaps, moving from data to a concept in one bound, are very risky; for the concept label, no matter how tempting a shortcut to expedite one’s analysis, one must take such leaps in baby steps: Develop and saturate your categories and move carefully developing attributes. Once you are certain about your data and the concepts, move it across the developed concept as if a template, to compare the fit, similarities, and comparisons of the two concepts.
But is this process always necessary? Our data are full of others’ concepts. Every sentence we utter has at least one established concept—that is how we communicate, for goodness sake. But in the example below, I describe how I found some very interesting data that could have been explained and better understood if I had placed it immediately within the context of a developed concept. Let us have a look.
The Experience of Agonizing Pain
At this point in my research program I was investigating how patients controlled agonizing pain, how they managed to maintain control of self. I had observed pain behavior in emergency and seen vocalizations and behaviors of the severely injured, and identified various states of discomfort from scared to out of control. I had conducted interviews, yet continued to feel that I had not reached the core of “what was going on.” One day while analyzing an interview from a lineman who had received a serious electrical burn, I walked away from my computer and came back to look at this text on the screen with new eyes:
… both hands and wrists were burnt, which were the [electrical] entry sites. And … both feet were burnt … the right side being the worst … right side of both hands and feet because—my right hand was up high—so naturally it entered the right the most … whatever was left over came down and went into the left … it still sustained some damage … the right hand had to be removed…. And the right leg was burnt quite badly, and the whole outside of the leg was burnt down to the bone—you could see bones and sinews in there, and at the very bottom—it also exited—not on the bottom of the foot—but … on the bottom of the leg, like right at the side of the foot was really burnt bad, too.
This person was talking about himself! Yet he was using depersonalized language. He was talking about himself as if his body were not his own. Let us code this piece of text using italics, so we can see the instances of depersonalization:
… both hands and wrists were burnt, which were the [electrical] entry sites. And … both feet were burnt … the right side being the worst … right side of both hands and feet because—my right hand was up high—so naturally it entered the right the most … whatever was left over came down and went into the left … it still sustained some damage … the right hand had to be removed…. And the right leg was burnt quite badly, and the whole outside of the leg was burnt down to the bone—you could see bones and sinews in there, and at the very bottom—it also exited—not on the bottom of the foot—but … on the bottom of the leg, like right at the side of the foot was really burnt bad, too. (Morse & Mitcham, 1998, p. 670)
The next step was to further analyze the text to see when the definitive article was used and not used. He said, “my right hand was up high—so naturally it entered the right the most” giving an instance of reverting to the possessive pronoun, perhaps indicting that before the burn, his body was still his.
An obvious conjecture could be: The way that persons with excruciating pain cognitively “manage” the pain, is to remove the pain from themselves (using “self” in the sociological sense), which is reflected in their language. In other words, when experiencing excruciating pain, processes of embodiment break down and painful body parts are disembodied. This disembodiment is reflected in their speech.
Therefore, are we seeing disembodiment? Can we immediately tag these data as disembodiment? Possibly.
Confirming the Phenomenon
At this moment in your analysis, the limitation is that we have only one case, and from that case made a cognitive leap from data to a concept. But with this new insight let us examine other cases from the data set for these “signals of disembodiment.”
First, we examined the descriptions of the pain experience of six other burn patients. All described the pain as “engulfing” and “overwhelming.” Four patients experienced amputations of digits, hands or feet or legs, and two had at least one prosthetic limb.
Therefore the next step was to conduct a secondary analysis of participants’ narratives to understand their experiences of illness, using data that were collected for previous studies but were not collected specifically to understand the pain experience. This point is important, as it removes any bias that may be with the interaction between the interviewer and the interviewee. If the interviews are already collected, the interviewee can’t possibly be responding to cues received from the interviewer. So we examined carefully the selected interviews that we decided to examine. Our rationale for the selection were:
1. Patients with spinal cord injury: Does the objectification of the body occur with loss of sensation and loss of control? Do patients who cannot feel parts of their body, treat those parts objectively?
2. Patients who had transplants (five heart, one heart–lung, one liver transplant, and one kidney transplant): Personal control and the amount of sensation of the body part remained unchanged before and after the transplant. Postoperative pain was controlled with analgesics.
3. Patients who had experienced myocardial infarction: These patients had no alteration in control of their hearts, nor loss of sensation. The patient experienced was less intense than burn patients.
Identification of Alternative Competing Explanatory Conjectures
The next step was to consider what this form of detached speech signifies. Our first two conjectures addressed the objectification of the self related to the transformation of body parts into a separated “object” or “thing” because the injury resulted in loss of sensation in the affected part or in the inability to control or to move the body part. These were:
1. When individuals lose control of a body part, some disembody that part to maintain control of the self.
2. When individuals lose sensation in a body part, disembodiment can be a psychological strategy for “disengaging” the body part from the self.
These two conjectures were tested by systematically comparing burn patient transcripts with patient groups who have and who have not lost sensation and control. Support for these conjectures is provided if the members of those populations that have lost sensation and/or control also used the definite article in reference to their bodies and if those who have not lost sensation or control do not use this form of speech.
3. Patients learn disembodying language from physicians.
Linguistically, modes of detachment are a normal pattern of “doctor talk” as a way of objectifying parts of the body. Patients learn these linguistic patterns of speech by hearing physicians discuss themselves as “cases” on rounds, and so forth.
4. Disembodiment is used in life-threatening trauma, in an effort to protect the self.
Here disembodiment would be viewed as a strategy encouraged by a threat to life in which removing the injured part of the self could remove the threat to the self. To support this conjecture, linguistic signals of disembodiment should occur in the transcripts during periods when patients face any life-threatening injury or illness.
5. Disembodiment is a strategy used to remove the body part, hence to remove the pain, when the agony is overwhelming.
This conjecture may be tested by comparing the burn patients with comparison groups that experienced varying amounts of pain in the course of their injury or illness.
At this point the presence or absence of each criterion is recorded for each conjecture for each patient group. It is important to note that the use of the criterion that a phenomenon be present in all cases during initial coding introduces an important standard of rigor required in qualitative work, especially considering the small samples used. But this need not mean that the appearance of one single negative case discounts or invalidates an observation. In quantitative inquiry a score of zero on a scale does not indicate that a phenomenon does not exist—only that it was not present in that particular case. One may thus suggest that these same standards could apply to qualitative inquiry. Yet in qualitative work, when the sample is obtained and analyzed case by case, if the negative case is the first case analyzed, there is a risk that this case could inadvertently redirect inquiry. The results are shown in Table 14.1.
From Table 14.1, the only conjecture clearly supported related to extreme pain. This supports the conclusion that disembodiment “disengages” a body part when a sensation is lost. Further examination of transcripts showed that burn patients tended to refer to their amputated limbs using the definitive article, “the leg” and to their prosthetic limbs using the possessive “my.” For example:
So [my wife] said, “Come up there with me.” And I started jumping about a foot, two feet in the air. I was doing fine until one leg pulled out and twisted sideways so I had to hang on to [my wife], and we had to … I sat down on the trampoline, and they had to get me off and put my leg back on … (laughter). So that was pretty hilarious. My brother was there, and he was just in tears laughing.