While working on a revision of the illustrations for the “Talking Shit in Rayne” essay, I considered for a time the use of a three-dimensional graph to represent how changes in genre interacted with changes in the interactional order. One axis was, of course, time. The second axis of the original graph was a kind of loose approximation of — okay a complete approximation of the degree of interaction occurring between the speaker and the audience.
In order not to have a fake 3D graph, of which there are plenty, I searched for a third axis of information that I could use. I fell upon the idea of trying to quantify the amount of interaction within a given text (semantic interaction, if you will) — as a way to explore the amount of interaction that a text created at the pragmatic level. My initial idea was to count the number of lines of direct discourse and then to divide by the total number of lines in a given text. That yielded the following results:
| Text
|
Direct Discourse Lines
|
Total Lines
|
Percentage
|
| 1.1
|
0
|
8
|
.55
|
| 1.2
|
6
|
11
|
.55
|
| 1.3
|
14
|
24
|
.58
|
| 1.4
|
12
|
14
|
.86
|
| 1.5
|
11
|
25
|
.44
|
| 1.6
|
25
|
34
|
.74
|
| 1.7
|
28
|
55
|
.51
|
| 1.8
|
41
|
81
|
.51
|
| 1.13
|
2
|
28
|
.07
|
| 1.14
|
20
|
52
|
.38
|
| 1.15
|
6
|
19
|
.32
|
| 1.17
|
8
|
21
|
.38
|
| 1.18
|
2
|
22
|
.09
|
A couple of notes first:
- Texts 1.1 – 1.4 are toasts.
- Texts 1.5 – 1.8 are jokes.
- Texts 1.13-1.18 are memorates, first-person accounts of encounters with the supernatural.
A couple of things stand out here:
- First, this kind of pattern seeking really wants more texts, and so there really isn’t enough data here.
- Second, I don’t even have all the data for all the texts I have — I was missing some transcripts when I did this.
- Third, despite all these qualifications, there are still some interesting patterns here:
- It’s hard to ignore the clustering of the first three texts, all of which are toasts, with values of .55-.58.
- It’s also hard to ignore the pair of jokes with a similar mirroring of percentages of direct discourse, despite enormous differences in length.
- The same can be said about the cluster of memorates, 14-17.
Another thing to note is that having to count lines of direct discourse was a bit more challenging than it seemed, though I think I made reliable distinctions. Just as importantly, however, was the fact that there were a few instances of indirect discourse that gave me pause. But very few. So few that now I want to go back and examine those instances more closely to see if I can’t discern any patterns of why it occurs at all.
