This repository has been archived by the owner on Jan 6, 2022. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
/
index.html
241 lines (188 loc) · 9.63 KB
/
index.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
<!DOCTYPE html>
<html lang=en>
<head>
<meta charset=utf-8>
<title>The cohesive function of prosody in Ékegusií (Kisii) folktales</title>
<meta name=author content='Daniel W. Hieber'>
<meta name=description content='Poster presented at the 48th Annual Conference on African Linguistics (ACAL), Indiana University, March 30 - April 2, 2017.'>
<meta name=keywords content='linguistics, language, functionalism, Kisii, Ékegusií, Bantu, East Africa, prosody, phonology, discourse'>
<meta name=viewport content='width=device-width, initial-scale=1'>
<link rel=stylesheet href=poster.css>
</head>
<body>
<header>
<h1>The cohesive function of prosody in Ékegusií (Kisii) folktales</h1>
<h2>Daniel W. Hieber (University of California, Santa Barbara) ~ <a href=http:https://danielhieber.com>danielhieber.com</a></h2>
<p>(click any spectrogram to play the associated audio)</p>
</header>
<main>
<section id=introduction class=section>
<h1 class=section-header>Introduction</h1>
<section class=section-content>
<section class=list-section>
<h2>Realizations of prosody</h2>
<ul>
<li>pause</li>
<li>accent</li>
<li>pitch reset</li>
<li>intonational parallelism</li>
<li>intonational contour</li>
</ul>
</section>
<section class=list-section>
<h2>Functions of Prosody</h2>
<ul>
<li>signals phonological structure</li>
<li>signals syntactic structure</li>
<li>signals discourse prominence</li>
<li>signals affect</li>
<li>abets information flow</li>
</ul>
</section>
<section>
<h2>Prosody as Cohesion</h2>
<ul>
<li>helps listeners to detect discourse boundaries, and the degree of cohesion or discontinuity between successive utterances</li>
<li>structure (whether phonological or syntactic) is always cohesion-forming</li>
</ul>
</section>
<section>
<strong>This study shows how prosodic features are used by speakers to lend cohesion to their discourse, by signaling the transitions from one unit of discourse to the next, the relations that hold between those units, and their relative prominence.</strong>
</section>
</section>
</section>
<!-- END INTRODUCTION -->
<section id=background class=section>
<h1 class=section-header>Background, Data & Methods</h1>
<section class=section-content>
<ul>
<li>Ékegusií (Gusii, Kisii), Great Lakes Bantu, southwest Kenya</li>
<li>Endangered: ~600,000 speakers, few under 30 y.o.</li>
<li>24 folktales recorded near Kisii Town, Kenya, Summer 2014</li>
<li>Lexical database, with audio, ~14,000 words</li>
<li>Texts annotated for prosodic features listed above</li>
<li>Prosodic boundaries were <em>not</em> marked - just the features</li>
</ul>
</section>
</section>
<!-- END BACKGROUND -->
<section id=pauses class=section>
<h1 class=section-header>Pauses</h1>
<section class=section-content>
<section>
<p>The length of pauses helps structure the text into its narrative stages</p>
</section>
<img src=img/pauses.jpg alt='chart showing length of each pause in a single text along a timeline'>
<section>
<ul>
<li>the participants are introduced and given descriptions, topic-comment style, each separated by a long pause</li>
<li>the complicating action is a sequence of utterances, where each utterance correspond to one step in a sequence of events</li>
<li>each step in the complicating action is followed by a long pause, creating a slower narrative pacing</li>
<li>pauses are extremely short during the movement toward the climax, creating a rapid pacing</li>
</ul>
</section>
<section>
<p><strong>Speakers use pauses to structure texts into major narrative sections, and control their pacing, thereby giving cohesion to each section.</strong></p>
</section>
</section>
</section>
<!-- END PAUSES -->
<section id=elision class=section>
<h1 class=section-header>Vowel Elision</h1>
<section>
<section class=section-content>
<ul>
<li>Ékegusií typically elides vowels at word junctions</li>
<li>Sometimes elision fails to occur phrase-internally, creating minor prosodic breaks</li>
</ul>
</section>
<section class=image-section>
<img class=spectrograph src=img/elision.png alt='spectrograph showing lack of vowel elision at the transition into reported speech'>
<audio src=audio/elision.wav controls preload></audio>
</section>
<section>
<p><strong>Vowel elision can be manipulated to create minor prosodic breaks even without pause, most typically at the transition to reported speech, or to demarcate distinct events in a rapid sequence.</strong></p>
</section>
</section>
</section>
<!-- END VOWEL ELISION -->
<section id=isotony class=section>
<h1 class=section-header>Intonational Parallelism</h1>
<section class=section-content>
<ul>
<li><dfn>Isotony</dfn> or <dfn>intonational parallelism</dfn> is the realization of the same intonational contour on separate, typically successive, stretches of speech.</li>
<li>By repeating the same intonational contour across two stretches of speech, the speaker highlights a similarity between them, whether in form or semantic content.</li>
</ul>
<section class=image-section>
<img class=spectrograph src=img/isotony.png alt='spectrograph illustrating isotony across multiple prosodic phrases'>
<audio src=audio/isotony.wav preload controls></audio>
</section>
</section>
</section>
<!-- END ISOTONY -->
<section id=reset class=section>
<h1 class=section-header>Pitch Reset</h1>
<section>
<section class=section-content>
<p>Problems determining pitch reset:</p>
<ul>
<li>canonical cases of pitch reset are rare in the Ékegusií corpus</li>
<li>terminal contours or prosodic accent may interrupt declination trends</li>
<li>pitch may reset down rather than up</li>
<li>many utterances show a flat contour or even a gradual rise</li>
<li>reported speech consistency shows a higher register</li>
</ul>
</section>
<section class=image-section>
<img class=spectrograph src=img/reset.png alt='spectrograph showing pitch reset over a series of utterances'>
<audio src=audio/reset.wav preload controls></audio>
</section>
<section>
<p><strong>Pitch reset is best understood as a more holistic property of how hearers perceive phrases, abstracting away from the parts of the pitch contour that are irrelevant to the overall trend.</strong></p>
</section>
</section>
</section>
<!-- END PITCH RESET -->
<section id=contour class=section>
<h1 class=section-header>Intonational Contour</h1>
<section class=section-content>
<p>Ékegusií has many types of intonational contours, each with a dedicated function. Classifying these contours as simply H or L, or falling or rising, fails to capture important differences between them. Some examples:</p>
<ul>
<li><strong>sharp terminal rise:</strong> backgrounding function, for conditionals, new topics, and ‘when’-clauses</li>
<li><strong>gradual terminal rise:</strong> continuing topic, more information to come</li>
<li><strong>high level register:</strong> extended sequences of events</li>
<li><strong>sharp fall to low register:</strong> conclusion of sequence of events</li>
</ul>
<p><strong>Speakers attend to the overall, conventionalized pattern of the intonational contour, rather than individual pitch targets.</strong></p>
</section>
</section>
<!-- END INTONATIONAL CONTOUR -->
<section id=conclusion class=section>
<h1 class=section-header>Conclusion</h1>
<section class=section-content>
<p>
Each prosodic feature helps delineate units of discourse and convey their relationship to the units around them. These features work in tandem to create stronger or weaker breaks in the discourse, thereby creating hierarchical prosodic structure.
</p>
</section>
</section>
</main>
<footer>
ACAL 48, Indiana University, March 30 - April 2, 2017. Research supported by NSF GRFP Grant No. 1144085.
</footer>
<script type=text/javascript>
const main = document.querySelector('main');
const spectrographs = new Map;
Array.from(document.querySelectorAll('.spectrograph')).forEach(graph => {
const audio = graph.nextElementSibling;
spectrographs.set(graph, audio);
});
const processClick = ev => {
if (spectrographs.has(ev.target)) {
const audio = spectrographs.get(ev.target);
audio.play();
}
};
main.addEventListener('click', processClick);
</script>
</body>
</html>