بازدید از آزمایشگاه

 
00:00 / 00:00
1.8x
1.4x
1.0x
0.7x
HD SD
HD
SD
اشتراک‌گذاری

×

گزارش خرابی

Hi, in the previous videos, we have seen
how linguists
can also work in the lab. How we can figure out
things about aphasia or about reading in the brain while
doing
experiments in a lab. So that's also a natural place
for a linguist to be. I'm going to discuss these issues
more with my student, Marten, now. >> So my main
question is, we've looked
at the results of what linguists do in the lab, but we haven't really
actually looked into the lab. So I was wondering, do you know how
these experiments
actually really work? >> Yeah, right, this is a good
question. Right so, so far,
although we say we are
doing experiments, we have been actually rather
theoretical about it. So we have been just talking
about
doing experiments rather than doing experiments
themselves. Now one reason for that is that the lab is not really the place
where I work as a
linguist a lot. But the good thing is that
fortunately,
here in Leiden, we do have linguistic
labs, and we have
talented young researchers working there. >> My name is Daan van der Velde. I'm a
linguist speech scientist
at Leiden University. I
research the speech of people who
are wearing a cochlear implant. A cochlear implant is a surgically
implanted hearing device for people who are deaf. >>
Uh-huh. >> So they will receive a surgery in
which a
wire with electrodes is inserted into their cochlea,
the cochlea are the inner part of the ear. That's why
it's called a cochlear implant. >> Right. >> And after that has been done, they can
to some extent
hear speech and music and all other sounds again. >> Right, so they don't need a hearing
device external
to the ear anymore because they have this thing-
>>
Precisely, yes. >> But then what does it sound like? Does it sound the same as the way we
perceive sound
now or is it different? >> It's quite different. They're trying to make it better and better but the way it works now,
the sound for them is very different. I can play an example for you. >> Yeah, please. >> This is a Dutch
sentence which has
been processed as if it sounds as if you are wearing a cochlear implant,
so it's a simulation. [SOUND] >> Right, so what I hear,
I know
nothing about this. What I hear is sort of patterns,
so
there is dynamic. But I can't really hear vowels or
consonants or stuff like that. >> That's right, yes. So, for instance, it's very hard for them to hear tones, melody and
speech, music is also very difficult. But you can hear some kind of rhythm,
so the dynamic
pattern and the temporal pattern, they are audible. So what we think is that the way they
listen to speech
is different from the way we listen to speech. They pay attention to different kinds of
aspects in
speech, different cues, and that's something that you
can research,
what they can hear, what they cannot hear. >> What sort of experiments do you do so
that you can research this? >> In this case, what we are doing is to
use simulation
of cochlear implants, so then you can use normal
hearing people and there are many more of those and
they're much more alike. That's very useful, and you
can manipulate
the way the speech is processed. With a
real cochlear implant,
you couldn't do that. It's just
the way an implant works for
specific patients. In this
case, we can manipulate things. But what we do here is
to simulate a
conversation between two normally hearing people and we process their speech as
if they are wearing a cochlear
implant. It's called faux coding. There's one speaker who is
in a recording booth, and there's a listener who
is in another booth. They will see the following
thing on the screen. First, there will be instructions. And
then they will see a sentence,
carrier sentence with four words. So they will have to pronounce
the sentence with the word that has been emphasized on the screen. >> Right, so the sentence is,
I say, and then a blank to you and they have to say one of these
four words on that blank. >> Yes, and the listener doesn't
know which word has
been pronounced. So the listener sees the same four
words and then has to choose, has to ask the speaker
is this
the word that you pronounced? Because it's not
clear for them,
it sounds very, very difficult. It's
very difficult to hear. So for them, it's not obvious
which word has been pronounced and then the speaker
will pronounce it again. And then we compare the
first
against the second speech segments. And then we
think that when they have to
emphasize their speech, they will do it in such a way that the listener can hear
it despite the distortion of
the speech. >> Yeah. >>
Right, so in this way,
you figure out how listening to language works while you're in this
very difficult circumstances. >> Yes. >> What is still important
about speech
of human language while you're already in these circumstances-
>> Yes, how you listen to speech. >> How do we listen to speech,
maybe a little more generally. >> Yes. >> Right, I guess this a good example why you would
go into the
lab in the first place. I mean how would you ever
figure out
what this speech sounds like for people who
have cochlear implants. You cannot go and
ask them
because they cannot tell you. You cannot try to do it
theoretically
because there's no way you can ever think
of that that sound you
just heard would be the actual sound. >> But surely there are more
experiments going on here, right? >> There must be more experiments. Maybe, well, we find another
talented young researcher here. >> Thank you. >> Hello.
>> Hi. >> What is your name? >> My name is Am Brasari. I am one of PhD
students in the Leiden
University Center for Linguistics. >> Good. What is your research? >> Mainly,
I'm working on field psycholinguistics. I want to know how
people actually produce, perceive languages. Basically, I'm working on second
language learners, how they produce non-letter vowels, vowels that did
not exist in their first
language. >> Right. >>
I want to see whether these vowels,
which are contrastive in their first language, that influence the process they
perceive and
produce second language. >> Right, because what I
would
think is that if, for example, there is a sound
in my language, like for
example in week two of the course, we talked about Hawaiian, and they don't
make a difference between t and
k. They're just the same sound. So if they would have to hear t and k, for example, so how does that work? >> Yeah,
what if second language learners
where I work in Javanese and Sundanese we don't have vowel the same as English. >> Right,
those are languages of Indonesia. >> Indonesians. >> Right.
>> We don't have
long vowels. We never say bid. We never say long E, long A, and so on. >> Right. >> Then would it be unfortunate for us to
know or to learn English or are we just okay, I don't
want to learn English
because I don't know the long vowels. No, of course not, right,
we have to find solution. We have to find why
actually that happened and one way to figure it out is well,
doing experiment. >> Right, so we do know, we do observe
that when we have Japanese
speakers, Sudanese speakers and
they're trying to
learn English, they have a hard time hearing
the
difference between tip and teep. >> Yeah. >> But now the question is,
what do they actually hear? And what do they actually say? Where is the problem? Because we might
want
to solve that problem. >> Interesting,
interesting. Well, we are actually working
on the
production of how they produce vowels in English and
we compare them with our control group. It's a bit
hard to do it really in the
real situation because we need to record them and we need to make them in the same
situation because this is
comparison, cross linguistic. So we compare the experimental
participants, Japanese and Sudanese, and we compare the result of
their speech production and perception with the native English. And we want to see how they are different. >> But what do you do? What do you do
with your participants? >> One thing I can show is I
did mouse tracking experiment. This is the experiment
which I
recorded their hand movement. >> Uh-huh,
because the mouse means a computer mouse. >> A
computer mouse,
not the mouse in the kitchen. >>
Right. >> Okay [LAUGH]
>> You have some software that
tracks how people move their mouse. >> True, so during our experiment, I will
ask them to sit down in
front of computer which has mouse tracking software. And it will appear the long and
short vowels into
different sides. Then they will click Start. >> Click me, please. >>
And then they have to move
their cursor right away, whether they hear bead or they hear bid. >> Right.
>>
Whether it's long vowels or short vowels. >> Right. >> So from
that moment,
if they directly choose the answer, it
makes smooth lines, or
whether they make a revision, they go negatively toward the incorrect
response and come up with the correct one. >> Right, so you bring your laptop to
Indonesia with
the mouse and then you do this experiment and the
software can track
exactly what people do with their mouse. So that you know whether they hesitate, maybe they go a little bit in the wrong
direction or they go straight. >> Whether they go a bit
slower than a native speaker, that's why we count on their reaction time
and also we see their
mouse trajectories. Of course,
computer mouse
trajectories, yeah. >> Right. >> Cool.
>> And so what kind of things do you find? Do you find that indeed people,
what kind of differences are there? >> Interestingly, yes, yes. >> Fortunately, yes. >> I compare the
result of control group,
native English speakers, of course, they directly make
a more direct execution. They directly answer that correctly,
smooth lines. Where all the second language learners, they make a kind of confusion
during their motor response. That's exactly true, yeah. >> That's very
interesting. >> Yeah.
>> Thank you, well, I guess this
is yet another example of why we need labs,
right? You can bring, can have a very small lab
which is just
the size of a laptop and a computer mouse, but
still
that counts as a little lab, and you find out things
which you would
never be able to find out otherwise. Now, in the next video, we asked also our informants
to do
tiny little experiments on themselves. They
didn't have labs but
we asked them to record some kind
of speech errors they
made in the past few weeks. And
we're going to analyze them,
together with the informants and also together with you.

دانلود با کیفیت بالا
دانلود با حجم کم