Prosodic and multimodal markers
of humor in conversation 41
2.2
Acoustical
analysis
A significant problem in doing acoustic analyses of naturally occurring conversa-
tions is the quality of the data. The instrumentation (see below for the technical
details) requires very “clean” recordings and any background noise degrades the
capacity of the software to perform the pitch analyses on which scholars base their
analyses. Recording high quality sound input often requires interfering with the
setup of the situation (for example by requiring the speakers to wear lavaliere mi-
crophones or to speak into a microphone). While this may be obviated by using
the recently introduced PZM (pressure zone) microphones, if one wants to have
both audio and video recording, the setup is prohibitively expensive and cumber-
some (multiple video cameras, synchronization of audio and video, etc.). Thus, we
chose a different solution to these problems.
We recorded a dyadic exchange in which each participant was seated in front
of a computer with a built-in camera and was wearing a headset/microphone com-
bination, of the type used commonly for chatting in online video-chats. This obvi-
ated the problem of audio and video recording, as well as providing a non-intrusive
means of recording the exchange, since the subjects were all reasonably familiar
with video chatting. Recordings were transferred to a Kay Pentax Computerized
Speech Laboratory (CSL) Model 4300b in order to make instrumental measures.
Pitch analysis was performed using the pitch contour analysis function of the CSL.
Volume was determined by performing an energy analysis using the energy con-
tour analysis function of the CSL. Pause length was calculated by selecting the area
of the sound wave surrounding the pause, performing a spectrographic analysis
and measuring the length of the pause manually using cursors. Speech rate was
calculated by dividing the overall time taken to produce the pause-based unit,
measured manually using cursors, by the number of syllables in the sample. Visual
inspection was used to identify smiling in the videos.
The conversation was collected at a small Midwestern university in the US.
Both the participants in the conversation were students who volunteered for the
data collection and signed an IRB release. All the participants knew one anoth-
er slightly from being enrolled in the same class. In this paper, we investigate a
conversation between two young women, Carmen and Marina (these are pseud-
onyms). The conversation lasted slightly over five minutes.
The participants were given a set of instructions: they were given the text of
the engineer and the frog joke, analyzed in Pickering et al. (2009), or another joke,
selected because it does not end with a dialogic punch line.
6
The participants were
6. We do not consider the results of the dialogic/non-dialogic punch lines in this article. They
will be the subject of ulterior publications.
42 Salvatore Attardo,
Lucy Pickering,
and Amanda Baker
told that they should tell one another their respective joke and that they should
continue talking for five or ten minutes, until told to stop by the person in charge
of the experiment. They were given the option of continuing to tell jokes or to talk
about anything they wanted. No further instructions were given.
The prosodic transcription conventions are as follows: following Brazil’s
(1997) model of discourse intonation and in the tradition of functionally based de-
scriptions of English intonation the conversations are divided in tone units. Tone
units boundaries are identified by the feature of prominence, i.e., fundamental
frequency (F0) peaks which distinguish prominent syllables from the surrounding
content. These prominent syllables are marked in CAPS (thus in line 1, the word
OK, has the second syllable capitalized, to indicate its prominence) and are con-
sidered to be the linguistically significant features of the tone unit are measured
for pitch and volume. Pitch and volume measurements for each prominent syllable
are provided, in that order. Figures are in Hertz and decibel, respectively, and are
in square brackets. A question mark indicates unreadable data. A range, such as
[232–204] on line 292, indicates a falling tone. Double forward slashes indicate
pauses and these pause-based units may comprise one or more tone unit. Pauses
are measured in seconds. All pauses are transcribed on a separate line, for ease
of identification. So, for example, line two has the first pause of the conversation,
which lasts 0.15 seconds (i.e., 15 hundredths of a second).
1
C
//UM [204][63] oKEY [225][69]//
2
0.15
3
C
// THERE [139][68] we GO [204][63] //
If a pause is followed by a notation in double round brackets, such as ((smack)),
this indicates that somewhere during the pause, there occurred a lip smacking
sound. Thus on line 166, we see a 2.5 seconds pause which includes a lip smacking,
within a turn by Carmen:
165
C
//YEAH [270][68] //
166
2.5 ((lip smack))
167
C
//YUP // [230] [60]
The numbering of the lines is for convenience of reference and has no theoretical
value. Author comments and impressionistic notes are in curly braces. Overlapping
turns are marked with a “=” in each speaker’s turn. The overlaps are not typo-
graphically marked by indentation. Other conversation analytical typographical
conventions are followed, such as marking lengthened vowels with colons.
Prosodic and multimodal markers of humor in conversation
43
2.3
A note on the coding
of humor and humor support
Generally, the production of humor to support humor is a well-documented phe-
nomenon. However, when coding a turn for analysis it is important to distinguish
between humorous turns of humor supports and non-humorous turns of humor
support. Consider the following exchange:
286
M
// you KNOW [147][65] = //
287
C
// =you’re aMERican [227][60] //
288
1.03
289
((both laugh))
290
M
// no KIdding [245][75] //
291
0.42
292
C
// YEAH [232–204] [71] (M. laughing) //
293
1.48
294
M
// DAMN [256][58] Americans //
295
0.09
296
C
//YESS [no Hz][43–71] //
On line 287, Carmen interrupts Marina with a jab, recycling the punch line of the
old joke about Americans being monolingual.
7
The joke is followed by a pause and
by shared laughter, which is a form of support (Hay 2001). Then Marina produces
an ironical supportive turn: “No kidding”. This is followed by a supportive turn
by Carmen “Yeah”. This turn is not humorous, while being supportive. Marina
7. What do you call someone who speaks three languages? Trilingual. What do you call some-
one who speaks two languages? Bilingual. How about someone who speaks only one language?
An American.
Chia sẻ với bạn bè của bạn: