1
00:00:37,670 --> 00:00:41,930
The introduction to this Unit
has already explained what respeaking is.

2
00:00:42,560 --> 00:00:46,050
The introduction
to this element has already explained

3
00:00:46,130 --> 00:00:49,970
what are the psycho-cognitive skills
a respeaker must possess.

4
00:00:50,040 --> 00:00:53,040
This video lecture is aimed
at letting you know

5
00:00:53,110 --> 00:00:55,680
the strategies a respeaker applies

6
00:00:55,750 --> 00:00:59,020
if he or she realises
the speaker is challenging.

7
00:00:59,080 --> 00:01:02,980
To do so, I will define the challenging speaker.

8
00:01:03,040 --> 00:01:06,970
 to then move to the strategies
a respeaker tries to apply

9
00:01:07,040 --> 00:01:10,400
when having to deal
with specific challenges

10
00:01:10,470 --> 00:01:12,680
like a fast speech,

11
00:01:12,740 --> 00:01:14,860
incomprehensible speech,

12
00:01:14,920 --> 00:01:16,700
a low-volume speech,

13
00:01:16,770 --> 00:01:18,620
an impromptu speech

14
00:01:18,680 --> 00:01:20,900
and plays on words.

15
00:01:22,310 --> 00:01:24,820
This is the agenda of this presentation.

16
00:01:32,880 --> 00:01:33,970
As you have seen,

17
00:01:34,030 --> 00:01:37,040
a respeaker has to do
many things at the same time.

18
00:01:37,860 --> 00:01:39,050
In this Element,

19
00:01:39,110 --> 00:01:42,080
we will deal with
the psychocognitive skill

20
00:01:42,150 --> 00:01:45,750
of activating exit strategies
while respeaking,

21
00:01:45,810 --> 00:01:48,190
if you realize
the speaker is challenging.

22
00:01:48,820 --> 00:01:49,740
In particular,

23
00:01:49,810 --> 00:01:53,440
for a respeaker to be able
and become a professional in the field,

24
00:01:53,500 --> 00:01:57,100
he or she has also to apply
solutions to problems

25
00:01:57,170 --> 00:01:59,640
he or she may encounter when subtitling,

26
00:01:59,710 --> 00:02:03,800
due to the fact
that he or she is no longer capable

27
00:02:03,900 --> 00:02:06,240
of providing the service required.

28
00:02:09,110 --> 00:02:12,150
While you can only develop
these competences by the time

29
00:02:12,210 --> 00:02:14,520
and thanks to ad hoc exercises,

30
00:02:14,590 --> 00:02:17,790
you may be interested in understanding
the rationale behind it.

31
00:02:18,420 --> 00:02:22,580
To do this, we will see
what a challenging speaker is.

32
00:02:23,070 --> 00:02:28,190
For every challenge,
I will try to provide concrete solutions.

33
00:02:31,090 --> 00:02:33,920
What does it mean
a speaker is challenging?

34
00:02:33,960 --> 00:02:36,960
There are many situations
where I found myself

35
00:02:37,060 --> 00:02:41,220
trying to subtitle a speech
I found it hard to subtitle.

36
00:02:41,580 --> 00:02:44,350
Here we will see five situations

37
00:02:44,420 --> 00:02:48,480
which may cause troubles to respeakers

38
00:02:48,540 --> 00:02:50,950
when subtitling a speech in real time.

39
00:02:51,470 --> 00:02:53,880
For sure, a speaker is challenging

40
00:02:53,950 --> 00:02:57,710
when he or she speaks above your MARS,

41
00:02:57,870 --> 00:03:02,770
meaning your capacity to produce text
as rapidly and accurately as needed,

42
00:03:02,800 --> 00:03:06,990
However, a fast speaker can be
less challenging than a speaker

43
00:03:07,060 --> 00:03:11,540
whose utterances are incomprehensible
for some reason

44
00:03:11,610 --> 00:03:16,100
or pronounced far from the mic
or at a low volume.

45
00:03:16,920 --> 00:03:19,140
These are to me the worst.

46
00:03:19,730 --> 00:03:22,140
Some colleagues also find it difficult

47
00:03:22,240 --> 00:03:26,590
to subtitle speakers who have not planned
their speech beforehand,

48
00:03:26,790 --> 00:03:29,960
and speakers who keep using
plays on words.

49
00:03:30,120 --> 00:03:34,180
Here are some examples
of challenging speakers

50
00:03:34,250 --> 00:03:39,590
and the possible solutions
to apply when you are in trouble.

51
00:03:46,820 --> 00:03:49,890
When trying to subtitle
people speaking very fast,

52
00:03:50,020 --> 00:03:52,070
there are three types of problems.

53
00:03:52,400 --> 00:03:57,110
One is that you cannot produce
accurate text at the same speed.

54
00:03:57,840 --> 00:04:01,600
A second problem is that you cannot
memorize or even understand

55
00:04:01,670 --> 00:04:03,320
what quick speakers say.

56
00:04:03,580 --> 00:04:04,870
Last but not least,

57
00:04:04,940 --> 00:04:09,650
even if you were capable of keeping
the same pace as the speaker,

58
00:04:09,720 --> 00:04:14,670
your audience would have problems
reading your subtitles at the same speed,

59
00:04:14,740 --> 00:04:20,680
especially if they are visible
as two-liner subtitles.

60
00:04:22,290 --> 00:04:25,160
So what to do
when a speaker speaks too fast?

61
00:04:26,020 --> 00:04:27,840
Let’s consider this example,

62
00:04:27,900 --> 00:04:31,530
which is taken
from the opening of the famous address

63
00:04:31,600 --> 00:04:35,430
delivered by Martin Luther King in 1963,

64
00:04:35,490 --> 00:04:38,360
at the March on Washington
for Jobs and Freedom.

65
00:04:38,730 --> 00:04:39,650
Let’s read it.

66
00:04:40,440 --> 00:04:43,150
I am happy to join with you today

67
00:04:43,220 --> 00:04:48,200
in what will go down in history
as the greatest demonstration for freedom

68
00:04:48,300 --> 00:04:50,200
in the history of our nation.

69
00:04:50,740 --> 00:04:53,640
Five score years ago, a great American,

70
00:04:53,710 --> 00:04:56,220
in whose symbolic shadow we stand today,

71
00:04:56,380 --> 00:04:58,730
signed the Emancipation Proclamation.

72
00:04:59,320 --> 00:05:03,150
This momentous decree came
as a great beacon light of hope

73
00:05:03,440 --> 00:05:06,840
to millions of Negro slaves
who had been seared

74
00:05:06,940 --> 00:05:09,380
in the flames of withering injustice.

75
00:05:10,110 --> 00:05:12,120
It came as a joyous daybreak

76
00:05:12,350 --> 00:05:15,030
to end the long night
of their captivity.

77
00:05:16,410 --> 00:05:21,560
Now let’s pretend we are subtitling
Dr. King’s speech in real time

78
00:05:21,630 --> 00:05:25,590
and that we were not given
the text of the speech beforehand.

79
00:05:26,020 --> 00:05:31,030
If we had the script of the speech,
we could feed the software with this

80
00:05:31,100 --> 00:05:33,240
or train ourselves with it.

81
00:05:33,970 --> 00:05:34,990
But we cannot.

82
00:05:35,420 --> 00:05:41,060
Let’s also pretend he is very stressed
and starts speaking very quickly.

83
00:05:41,790 --> 00:05:42,950
What can we do?

84
00:05:44,930 --> 00:05:49,120
One solution may be
the Gordian Knot Strategy,

85
00:05:49,190 --> 00:05:50,680
or cut and knit.

86
00:05:50,740 --> 00:05:52,460
It means that you can subtitle

87
00:05:52,520 --> 00:05:55,560
only half of the sentences
pronounced by the speaker

88
00:05:55,620 --> 00:05:57,770
while trying to keep coherence.

89
00:05:58,100 --> 00:05:59,020
In this example,

90
00:05:59,190 --> 00:06:02,060
you can avoid pronouncing
every other sentence

91
00:06:02,220 --> 00:06:04,700
and still provide
understandable subtitles,

92
00:06:04,770 --> 00:06:06,810
with little or no cohesion effort.

93
00:06:07,270 --> 00:06:10,440
Let’s read the text
as a result of this strategy.

94
00:06:10,870 --> 00:06:14,430
We will see it keeps saying
more or less the same thing.

95
00:06:16,180 --> 00:06:18,230
I am happy to join with you today.

96
00:06:18,690 --> 00:06:20,180
Five score years ago,

97
00:06:20,340 --> 00:06:23,870
a great American signed
the Emancipation Proclamation.

98
00:06:24,330 --> 00:06:28,100
This momentous decree came
as a great beacon light of hope

99
00:06:28,360 --> 00:06:30,100
to millions of Negro slaves.

100
00:06:30,640 --> 00:06:35,290
It came as a joyous daybreak
to end the long night of their captivity.

101
00:06:38,210 --> 00:06:40,420
Of course this is not always easy.

102
00:06:40,480 --> 00:06:43,590
Sometimes, doing this may cause
even more troubles.

103
00:06:44,110 --> 00:06:46,490
Let’s consider the same example,

104
00:06:46,750 --> 00:06:51,140
and apply the Gordian Knot strategy
in a different manner.

105
00:06:51,700 --> 00:06:56,130
Let’s read the text as a result
of a badly applied Gordian Knot strategy.

106
00:06:57,550 --> 00:07:00,190
I am happy to join with you today

107
00:07:00,550 --> 00:07:05,040
in what will go down in history
as the greatest demonstration for freedom

108
00:07:05,300 --> 00:07:07,040
in the history of our nation.

109
00:07:07,610 --> 00:07:12,230
This momentous decree came
as a great beacon light of hope

110
00:07:12,360 --> 00:07:14,240
to millions of Negro slaves,

111
00:07:14,410 --> 00:07:17,840
who had been seared
in the flames of withering injustice.

112
00:07:18,470 --> 00:07:22,590
But one hundred years later,
the Negro still is not free.

113
00:07:25,560 --> 00:07:27,610
As we can see,

114
00:07:28,630 --> 00:07:34,340
the application of the strategy creates
some inconsistencies in this case.

115
00:07:34,930 --> 00:07:37,310
We have left out the sentence

116
00:07:37,340 --> 00:07:41,440
with Abraham Lincoln having signed
the Emancipation Proclamation.

117
00:07:43,050 --> 00:07:47,310
This creates a huge cohesion problem,

118
00:07:47,380 --> 00:07:50,180
because the following sentences
refer to it.

119
00:07:52,290 --> 00:07:57,740
Are we still capable of guessing
what the speaker means by "Negro slaves"?

120
00:08:00,670 --> 00:08:06,650
Is he referring to the same people
the speaker is addressing to?

121
00:08:06,780 --> 00:08:10,010
Or to the ones he's really referring to?

122
00:08:11,000 --> 00:08:14,500
And what does
the "hundred years later" phrase refer to?

123
00:08:18,080 --> 00:08:21,580
When you are not confident
going for the Gordian Knot,

124
00:08:21,970 --> 00:08:24,080
prefer the Garwood Strategy.

125
00:08:24,550 --> 00:08:26,160
When in doubt, leave it out.

126
00:08:26,720 --> 00:08:28,700
You leave out an entire chunk.

127
00:08:29,030 --> 00:08:31,050
Then, start from the first sentence

128
00:08:31,110 --> 00:08:33,420
it makes sense to re-start from.

129
00:08:34,010 --> 00:08:36,520
In this case, it is the last sentence

130
00:08:36,690 --> 00:08:39,300
without the reference
to that "something",

131
00:08:39,370 --> 00:08:43,450
you were not able to subtitle
which happened 100 years before.

132
00:08:43,880 --> 00:08:45,730
The final result is the following.

133
00:08:46,560 --> 00:08:48,570
I am happy to join with you today

134
00:08:48,800 --> 00:08:53,650
in what will go down in history
as the greatest demonstration for freedom

135
00:08:53,980 --> 00:08:55,650
in the history of our nation.

136
00:08:56,590 --> 00:08:58,630
The Negro still is not free.

137
00:09:00,320 --> 00:09:04,240
This is an example
of an extreme solution.

138
00:09:04,710 --> 00:09:08,200
The audience realizes
that subtitles do not read

139
00:09:08,500 --> 00:09:10,050
what is said by the speaker.

140
00:09:10,910 --> 00:09:16,160
This is not a problem
if this is rarely adopted

141
00:09:16,220 --> 00:09:21,900
and if you are capable of explaining
the reason why it happened.

142
00:09:29,760 --> 00:09:34,610
Do you remember that famous scene
where Charles Chaplin in the Great Dictator

143
00:09:35,430 --> 00:09:38,630
speaks fake German
to the people of Tomainia.

144
00:09:38,730 --> 00:09:40,320
who applauds him?

145
00:09:40,810 --> 00:09:47,120
When we respeak, it may happen
that we as real-time subtitlers

146
00:09:47,180 --> 00:09:49,560
find ourselves in situations

147
00:09:49,630 --> 00:09:53,150
where we don’t understand
one or more sentences

148
00:09:53,350 --> 00:09:55,230
that the speaker pronounces.

149
00:09:55,270 --> 00:09:58,930
However, it seems
that his or her audience does understand,

150
00:09:58,930 --> 00:10:01,140
as it happens in this famous scene.

151
00:10:02,370 --> 00:10:06,200
In such cases,
you can adopt the Generalisation Strategy.

152
00:10:06,550 --> 00:10:08,840
If you know the topic and the context,

153
00:10:09,000 --> 00:10:12,990
and you more or less know
what he or she is saying,

154
00:10:13,090 --> 00:10:17,450
you can relate
to the last concept you have understood.

155
00:10:18,040 --> 00:10:20,320
You can either reformulate it,

156
00:10:20,980 --> 00:10:25,440
or say something
that does not commit yourselves too much,

157
00:10:25,500 --> 00:10:27,550
like something general or logical.

158
00:10:28,080 --> 00:10:29,890
Let’s consider some examples.

159
00:10:30,640 --> 00:10:33,120
Let’s consider
the same example as above.

160
00:10:33,590 --> 00:10:36,920
You stop understanding the speaker
when he says:

161
00:10:37,250 --> 00:10:40,880
“This momentous decree came
as a great beacon light of hope

162
00:10:41,040 --> 00:10:42,890
to millions of Negro slaves,

163
00:10:43,090 --> 00:10:47,550
who had been seared
in the flames of withering injustice”.

164
00:10:48,540 --> 00:10:49,530
What can you do?

165
00:10:50,710 --> 00:10:54,510
You can try to keep providing
the subtitling service,

166
00:10:54,840 --> 00:10:58,900
by reformulating
the last concept you have understood,

167
00:10:59,290 --> 00:11:03,190
meaning: "Five score years ago,

168
00:11:03,220 --> 00:11:06,690
a great American,
in whose symbolic shadow we stand today,

169
00:11:06,850 --> 00:11:09,520
signed the Emancipation Proclamation".

170
00:11:11,040 --> 00:11:12,160
Here is the result.

171
00:11:12,590 --> 00:11:14,540
The sentence is replaced

172
00:11:14,670 --> 00:11:18,000
by the reformulation
of the last subtitled sentence,

173
00:11:18,170 --> 00:11:22,820
meaning: "The Emancipation proclamation
signed 100 years ago,

174
00:11:23,190 --> 00:11:27,410
by such a Great American,
is the symbol of us all today".

175
00:11:28,290 --> 00:11:31,770
It may sound a bit redundant
with the previous sentence,

176
00:11:31,860 --> 00:11:35,730
but this is an inspirational speech
with an important use of rhetoric.

177
00:11:35,990 --> 00:11:38,560
Moreover, subtitles appear in two lines.

178
00:11:38,630 --> 00:11:41,140
So,
the previous sentence has already gone.

179
00:11:43,880 --> 00:11:47,280
This strategy is not to be confused
with cheating.

180
00:11:47,840 --> 00:11:49,280
Here you don’t lie.

181
00:11:49,310 --> 00:11:52,520
On the contrary,
you keep providing correct information.

182
00:11:52,920 --> 00:11:56,420
Consequently, the audience keeps relying
on the subtitles

183
00:11:56,610 --> 00:11:59,950
because they keep seeing them
under the speaker’s mouth.

184
00:12:02,190 --> 00:12:03,780
Let's consider another case.

185
00:12:04,170 --> 00:12:07,570
Let's pretend you understand
the overall message

186
00:12:07,650 --> 00:12:09,390
the speaker wants to deliver

187
00:12:09,650 --> 00:12:15,660
thanks to context, body language
or from slides he or she may use.

188
00:12:17,340 --> 00:12:19,880
You can replace the missing bits

189
00:12:19,980 --> 00:12:23,280
with logical consequences
or general statements,

190
00:12:23,340 --> 00:12:26,940
without jeopardising
understanding of the source text,

191
00:12:27,470 --> 00:12:30,970
or cheating with fake information,

192
00:12:31,330 --> 00:12:36,180
or repeating the same information,
as in the previous example.

193
00:12:36,640 --> 00:12:38,620
The final result is the following.

194
00:12:39,350 --> 00:12:42,750
"This was a fundamental step
in our history

195
00:12:43,140 --> 00:12:45,850
that made this gathering possible".

196
00:12:47,130 --> 00:12:48,580
Here you have replaced

197
00:12:48,650 --> 00:12:51,890
the incomprehensible sentence
with a general statement,

198
00:12:51,950 --> 00:12:55,520
which is also a logical continuation
of the previous sentence.

199
00:12:56,400 --> 00:13:01,160
When you are no confident going
for the Generalisation Strategy, again,

200
00:13:01,220 --> 00:13:04,790
prefer the Garwood Strategy:
when in doubt, leave it out.

201
00:13:05,320 --> 00:13:07,960
Remember that this is an extreme solution

202
00:13:08,030 --> 00:13:11,820
and that the audience realizes
that subtitles do not read

203
00:13:11,890 --> 00:13:13,730
what is said by the speaker.

204
00:13:20,630 --> 00:13:26,040
Some speakers, like Marlon Brando's
Don Corleone in The Godfather

205
00:13:26,100 --> 00:13:29,960
who mumble or speak with a low volume.

206
00:13:30,030 --> 00:13:33,030
The problem is not
that you don’t understand them.

207
00:13:33,590 --> 00:13:35,770
With a little effort you do.

208
00:13:36,140 --> 00:13:39,200
The problem is when you try
to respeak them.

209
00:13:39,340 --> 00:13:42,540
You no longer manage to listen
and speak at the same time,

210
00:13:42,710 --> 00:13:46,370
because your voice covers
that of the speaker.

211
00:13:46,660 --> 00:13:50,820
Moreover, the extra effort involved
in understanding the speaker,

212
00:13:50,890 --> 00:13:53,690
reduces your capacity to respeak.

213
00:13:53,920 --> 00:13:56,200
So, what can we do in such cases?

214
00:13:58,110 --> 00:14:02,210
If the speaker speaks out of the mike
or with a low volume,

215
00:14:02,630 --> 00:14:04,380
there is little we can do.

216
00:14:04,520 --> 00:14:08,310
You can ask the speaker
to speak closer to the mic or louder,

217
00:14:08,440 --> 00:14:11,050
if you are physically close
to him or her.

218
00:14:11,410 --> 00:14:13,920
However, sometimes it happens

219
00:14:13,990 --> 00:14:18,380
that a speaker used to do this,
speaking with a low volume

220
00:14:18,440 --> 00:14:20,690
or not caring about the microphone,

221
00:14:22,730 --> 00:14:26,230
well, he or she will not manage
to behave for long.

222
00:14:26,960 --> 00:14:29,200
In such cases, you can only adopt

223
00:14:29,230 --> 00:14:32,700
the Gordian Knot strategy,
or cut and knit strategy.

224
00:14:33,340 --> 00:14:34,660
As seen previously,

225
00:14:34,860 --> 00:14:39,250
the Gordian Knot consists
in hearing a sentence

226
00:14:39,380 --> 00:14:42,980
and repeating it as fast as possible
once it is over,

227
00:14:43,270 --> 00:14:45,250
by compressing it if you can.

228
00:14:46,280 --> 00:14:48,920
Then, you skip the following sentence.

229
00:14:48,980 --> 00:14:51,660
Concentrate
on the third sentence of the row.

230
00:14:52,280 --> 00:14:57,200
You are going to repeat it
as quickly as possible once it is over,

231
00:14:57,500 --> 00:15:02,680
by paying attention to cohesion
and coherence with the first sentence.

232
00:15:03,500 --> 00:15:05,420
As in the case of a fast speaker,

233
00:15:05,520 --> 00:15:08,400
this is something
that you cannot do for long.

234
00:15:15,320 --> 00:15:19,130
There is a huge debate,
among scholars and among professionals,

235
00:15:19,260 --> 00:15:24,710
about whether one can subtitle
all features of orality in a speech.

236
00:15:25,500 --> 00:15:26,980
Somebody says yes,

237
00:15:27,050 --> 00:15:31,070
because one must be faithful
to what the speaker actually says.

238
00:15:31,290 --> 00:15:32,540
Somebody says no,

239
00:15:32,810 --> 00:15:37,080
because readability
and understandability is more important.

240
00:15:37,770 --> 00:15:40,080
Especially when it comes to a politician,

241
00:15:40,280 --> 00:15:44,640
the question of reproducing
one's bad grammar or mistakes

242
00:15:44,740 --> 00:15:46,640
is all the more important.

243
00:15:46,810 --> 00:15:50,410
It has to do with
being fair with the opposition,

244
00:15:50,540 --> 00:15:52,620
or diplomatic with the speaker.

245
00:15:53,320 --> 00:15:56,150
Though an answer is difficult to provide

246
00:15:56,320 --> 00:15:59,190
because, to me,
both parties have their point,

247
00:15:59,620 --> 00:16:03,810
Let me now concentrate
on how to deal with features of orality.

248
00:16:04,700 --> 00:16:10,510
They jeopardise both your understanding
as a real-time subtitler

249
00:16:10,640 --> 00:16:14,070
and that of the audience reading
verbatim subtitles.

250
00:16:14,300 --> 00:16:17,040
So, what can we do?

251
00:16:17,800 --> 00:16:22,030
Though we always have to understand
which strategy is better to use,

252
00:16:22,190 --> 00:16:24,110
verbatim or sensatim,

253
00:16:24,240 --> 00:16:28,130
an impromptu speech is an obstacle
to the subtitling process,

254
00:16:28,260 --> 00:16:33,380
especially in case respeaking
is used as a subtitle tool.

255
00:16:33,510 --> 00:16:35,920
Respeaking is not that fit

256
00:16:36,050 --> 00:16:38,860
to reproduce non-existing words,

257
00:16:38,990 --> 00:16:41,630
like those speakers,
especially foreigners

258
00:16:43,010 --> 00:16:47,270
may produce as a result of calques
from their native language

259
00:16:47,470 --> 00:16:50,010
or as a result of mispronunciation.

260
00:16:50,543 --> 00:16:53,540
So, how to cope with impromptu speech?

261
00:16:54,330 --> 00:16:57,341
If it is a challenge
to the subtitling process,

262
00:16:57,539 --> 00:17:00,278
you can go for correcting
one’s bad grammar

263
00:17:00,377 --> 00:17:02,291
and go straight to the point.

264
00:17:02,654 --> 00:17:06,317
If it is not relevant
that the speaker makes mistakes

265
00:17:06,416 --> 00:17:11,531
because he is a non-native,
speaking English as a communication tool,

266
00:17:11,894 --> 00:17:15,590
no matter if he or she
doesn’t pronounce words properly,

267
00:17:15,854 --> 00:17:18,956
selects wrong words, mumbles,

268
00:17:19,055 --> 00:17:23,840
hesitates, self-reformulates,
produces extra sounds and so on.

269
00:17:24,104 --> 00:17:28,394
What is important in a conference speech
is the main message.

270
00:17:29,087 --> 00:17:32,651
If you manage
to understand what the speaker says,

271
00:17:32,651 --> 00:17:37,667
and recognize pronunciation,
grammar and lexical mistakes,

272
00:17:37,766 --> 00:17:40,076
due to the speaker’s native language,

273
00:17:40,208 --> 00:17:44,498
you can go and correct
the speaker’s output.

274
00:17:45,158 --> 00:17:47,798
Preferably, with short sentences.

275
00:17:48,029 --> 00:17:52,583
Since foreigners tend to speak slowly
if compared to native speakers,

276
00:17:52,781 --> 00:17:57,335
the strategy is to wait
for a complete sentence and then respeak.

277
00:17:58,160 --> 00:18:02,252
The delay should be enough
to let you produce correct sentences.

278
00:18:02,945 --> 00:18:05,750
Don’t try to follow
the pace of the speakers

279
00:18:05,882 --> 00:18:11,525
Because of non-natural pauses,
the software may cause recognition errors

280
00:18:11,723 --> 00:18:17,795
which are best avoided by pronouncing
sentences as naturally as possible.

281
00:18:19,874 --> 00:18:23,504
Concretely, let’s consider
this real-life example

282
00:18:23,603 --> 00:18:28,355
taken from the transcription
of a speech made by an Italian politician

283
00:18:28,553 --> 00:18:32,150
famous for not being
an excellent speaker of English.

284
00:18:33,173 --> 00:18:37,595
[reads the slide]

285
00:19:04,490 --> 00:19:07,196
This speaker is challenging.

286
00:19:07,196 --> 00:19:13,103
especially if you don't guess
the message the speaker wants to deliver.

287
00:19:13,235 --> 00:19:19,140
or if you don't understand
the way the speaker pronounces words.

288
00:19:19,901 --> 00:19:26,171
However, truncating what the speaker says
into smaller chunks is quite easy

289
00:19:26,402 --> 00:19:30,692
because he makes use of a simple grammar.

290
00:19:34,718 --> 00:19:37,622
Considering the examples we have just seen

291
00:19:38,612 --> 00:19:45,278
the speaker is challenging
if you don't guess hwat he means.

292
00:19:45,443 --> 00:19:51,647
However, truncating what the speaker says
into smaller chunks is quite easy

293
00:19:52,010 --> 00:19:56,795
because the speaker, like many foreigners,
uses a simple grammar.

294
00:19:56,894 --> 00:19:59,666
He also makes quite many pauses.

295
00:20:00,128 --> 00:20:04,682
The possible result
of the application of an exit strategy

296
00:20:04,913 --> 00:20:08,048
which is that of producing short sentences

297
00:20:08,774 --> 00:20:13,889
is to keep
as much of the original as possible.

298
00:20:16,001 --> 00:20:20,819
[he reads the slide]

299
00:20:40,949 --> 00:20:43,490
OK, when doing this, remember this.

300
00:20:43,589 --> 00:20:48,044
This manipulation of the Source Text
is only possible at the grammar level.

301
00:20:48,143 --> 00:20:50,024
Not at the content level.

302
00:20:50,453 --> 00:20:55,205
Otherwise this may cause differences
between those who hear the speaker

303
00:20:55,304 --> 00:20:57,482
and those who read the subtitles.

304
00:20:58,175 --> 00:21:02,366
Also consider that somebody
having heard what the speaker says

305
00:21:02,696 --> 00:21:05,237
may contradict him or her

306
00:21:05,336 --> 00:21:07,679
or correct him or her.

307
00:21:08,042 --> 00:21:15,434
So it is important subtitles
are consistent with what the speaker says.

308
00:21:24,010 --> 00:21:28,271
Plays on words are really terrible
to respeak for many reasons.

309
00:21:28,370 --> 00:21:31,241
especially if you have to subtitle them live.

310
00:21:32,198 --> 00:21:36,686
In the movie Pulp Fiction,
Uma Thurman tells a joke to John Travolta

311
00:21:36,785 --> 00:21:41,780
[he reads the slide]

312
00:21:53,516 --> 00:21:56,030
I don’t want to know
how much time it took

313
00:21:56,130 --> 00:21:59,330
to those who had to subtitle this
into other languages,

314
00:21:59,400 --> 00:22:01,480
or if they managed to do it.

315
00:22:02,110 --> 00:22:06,000
But even intralingually,
this is not an easy task.

316
00:22:06,122 --> 00:22:08,762
This joke is based on two homophones.

317
00:22:08,960 --> 00:22:12,095
The verb "catch up"
as a synonym to "speed up"

318
00:22:12,227 --> 00:22:16,055
and the noun "ketchup",
the tomato-based sauce.

319
00:22:17,540 --> 00:22:21,335
When it is a play involving
a non-standard spelling,

320
00:22:21,566 --> 00:22:26,912
we normally stop respeaking
and type the word in uppercase letters.

321
00:22:27,044 --> 00:22:31,664
 It is a signal to the audience
that that is the word or group of words

322
00:22:31,961 --> 00:22:35,063
where they have to focus their attention,

323
00:22:35,162 --> 00:22:37,769
as that is the funny part of the joke.

324
00:22:37,835 --> 00:22:40,772
 Some colleagues use inverted commas.

325
00:22:40,904 --> 00:22:44,798
Some others put an exclamation mark
after the wordplay.

326
00:22:44,930 --> 00:22:48,725
Some others use
the caption [laughter].

327
00:22:49,649 --> 00:22:52,256
What do we do in this case,

328
00:22:52,355 --> 00:22:55,325
where a homophone is used
to provoke the laughter?

329
00:22:56,920 --> 00:23:02,057
One solution can be that
of spelling the non-standard homophone,

330
00:23:02,453 --> 00:23:04,235
the unexpected one.

331
00:23:04,490 --> 00:23:08,250
In this case we use ellipses
as a suspense technique.

332
00:23:08,620 --> 00:23:10,660
Then we use uppercase letters

333
00:23:10,730 --> 00:23:14,230
to say that they have to concentrate
on the word Ketchup,

334
00:23:14,890 --> 00:23:19,740
the unexpected, laughter-provoking synonym
to catch up or speed up.

335
00:23:25,685 --> 00:23:29,100
In this video lecture,
we have seen some concrete examples

336
00:23:29,100 --> 00:23:32,130
of how to behave
when a speaker is challenging.

337
00:23:32,830 --> 00:23:34,340
When a speaker is challenging,

338
00:23:34,640 --> 00:23:38,300
we need to adopt
what we call here Exit Strategies,

339
00:23:38,500 --> 00:23:43,420
meaning strategies to apply to exit
a potentially dangerous situation,

340
00:23:43,780 --> 00:23:46,920
for either the subtitler,
the audience or both.

341
00:23:47,770 --> 00:23:49,360
In case of a fast speaker,

342
00:23:49,560 --> 00:23:52,360
the mostly used technique
is the Gordian Knot,

343
00:23:52,690 --> 00:23:54,410
or cut and knit strategy.

344
00:23:54,800 --> 00:23:59,920
You try to reduce the speech
by omitting whole chunks

345
00:24:00,050 --> 00:24:05,000
and managing to make the non-omitted parts
coherent among them.

346
00:24:05,890 --> 00:24:08,600
An extreme solution
is the Garwood strategy,

347
00:24:08,930 --> 00:24:13,050
meaning leaving everything out
and explaining why you did so.

348
00:24:14,110 --> 00:24:17,180
The Garwood strategy
is also an extreme strategy

349
00:24:17,540 --> 00:24:19,620
for incomprehensible speeches.

350
00:24:20,020 --> 00:24:25,460
However, replacing incomprehensible bits
with general or logical bits

351
00:24:25,460 --> 00:24:26,850
is to be preferred.

352
00:24:27,870 --> 00:24:32,690
When a speaker speaks with a low volume,
the Gordian Knot is the best solution.

353
00:24:33,350 --> 00:24:39,250
When a speaker makes an Impromptu speech
with mistakes and features of orality,

354
00:24:39,620 --> 00:24:41,860
the strategy is to skip them

355
00:24:42,220 --> 00:24:45,850
and try to produce
short and correct sentences.

356
00:24:46,580 --> 00:24:49,910
Finally,
when it comes to plays on words,

357
00:24:49,910 --> 00:24:54,340
the best solution is to type
the laughter-provoking unexpected word

358
00:24:54,860 --> 00:24:56,810
in uppercase characters.