WEBVTT

00:00:00.000 --> 00:00:01.944 align:middle line:90%
[SQUEAKING]

00:00:01.944 --> 00:00:03.402 align:middle line:90%
[RUSTLING]

00:00:03.402 --> 00:00:06.804 align:middle line:90%
[CLICKING]

00:00:06.804 --> 00:00:13.120 align:middle line:90%


00:00:13.120 --> 00:00:16.410 align:middle line:84%
NORVIN RICHARDS: Today, we
are going to finish syntax,

00:00:16.410 --> 00:00:19.390 align:middle line:84%
not in the sense that you
will now know everything

00:00:19.390 --> 00:00:20.640 align:middle line:90%
there is to know about syntax.

00:00:20.640 --> 00:00:24.150 align:middle line:84%
People spend their
lives studying syntax.

00:00:24.150 --> 00:00:27.000 align:middle line:90%
In fact, that's what I'm doing.

00:00:27.000 --> 00:00:29.790 align:middle line:84%
But you'll know
enough for 24.900.

00:00:29.790 --> 00:00:31.830 align:middle line:84%
And if you're interested
in these things,

00:00:31.830 --> 00:00:35.190 align:middle line:84%
there is 24.902, which
is all about syntax--

00:00:35.190 --> 00:00:38.910 align:middle line:84%
introduction-- more of an
introduction to syntax.

00:00:38.910 --> 00:00:40.890 align:middle line:84%
So I was in the middle
of talking, last time,

00:00:40.890 --> 00:00:44.370 align:middle line:84%
about shortest move, and so I'm
going to back up a little bit

00:00:44.370 --> 00:00:46.050 align:middle line:84%
and we'll talk through
the cases that I

00:00:46.050 --> 00:00:49.130 align:middle line:90%
was talking about last time.

00:00:49.130 --> 00:00:50.880 align:middle line:84%
Before I do that,
though, I wanted to say,

00:00:50.880 --> 00:00:54.720 align:middle line:84%
I got a really interesting
question after class last time

00:00:54.720 --> 00:00:57.570 align:middle line:84%
from someone who basically
wanted to know whether it could

00:00:57.570 --> 00:01:02.310 align:middle line:84%
be shown that the kinds of
universals that I was showing

00:01:02.310 --> 00:01:06.600 align:middle line:84%
you, observations like there's
wh- movement to the left,

00:01:06.600 --> 00:01:10.260 align:middle line:84%
but there isn't wh- movement
to the right, or there's V2,

00:01:10.260 --> 00:01:12.173 align:middle line:90%
but there isn't anti-V2--

00:01:12.173 --> 00:01:13.590 align:middle line:84%
Somebody wanted
to know whether it

00:01:13.590 --> 00:01:15.548 align:middle line:84%
could be shown that the
absence of these things

00:01:15.548 --> 00:01:16.810 align:middle line:90%
was statistically significant.

00:01:16.810 --> 00:01:19.740 align:middle line:84%
That is, if you look at
the languages of the world,

00:01:19.740 --> 00:01:23.580 align:middle line:84%
would we expect to find
anti-V2 is V2 common enough

00:01:23.580 --> 00:01:27.130 align:middle line:84%
that it's surprising that
the anti-V2 is absent.

00:01:27.130 --> 00:01:30.480 align:middle line:84%
So I am working on
working up some stats,

00:01:30.480 --> 00:01:32.820 align:middle line:84%
and I will probably
send out an announcement

00:01:32.820 --> 00:01:33.970 align:middle line:90%
about that stuff later.

00:01:33.970 --> 00:01:37.950 align:middle line:84%
So keep your eyes
peeled for that.

00:01:37.950 --> 00:01:41.830 align:middle line:84%
In working up some stats, I'm
having a look at a website,

00:01:41.830 --> 00:01:43.585 align:middle line:90%
which I can recommend.

00:01:43.585 --> 00:01:45.960 align:middle line:84%
It's called WALS-- oh dear,
I was about to write the URL,

00:01:45.960 --> 00:01:47.502 align:middle line:84%
and then I realized
I don't know how.

00:01:47.502 --> 00:01:49.260 align:middle line:90%
If you Google WALS--

00:01:49.260 --> 00:01:56.653 align:middle line:84%
WALS stands for the World
Atlas of Language Structures.

00:01:56.653 --> 00:01:59.070 align:middle line:84%
And I think it's something
dumb like WALS.com or something

00:01:59.070 --> 00:02:01.392 align:middle line:90%
like that.

00:02:01.392 --> 00:02:02.710 align:middle line:90%
AUDIENCE: [INAUDIBLE]

00:02:02.710 --> 00:02:04.035 align:middle line:90%
NORVIN RICHARDS: WALS dot what?

00:02:04.035 --> 00:02:04.827 align:middle line:90%
AUDIENCE: Dot info.

00:02:04.827 --> 00:02:06.000 align:middle line:90%
NORVIN RICHARDS: Info, huh?

00:02:06.000 --> 00:02:09.150 align:middle line:90%
No wonder I don't know the URL.

00:02:09.150 --> 00:02:11.650 align:middle line:90%
Language Structures-- Thank you.

00:02:11.650 --> 00:02:17.580 align:middle line:84%
It is a typological
database, gathering

00:02:17.580 --> 00:02:20.860 align:middle line:84%
typological information about
a bunch of different languages.

00:02:20.860 --> 00:02:25.620 align:middle line:84%
So it has thousands of
languages, observations

00:02:25.620 --> 00:02:28.830 align:middle line:84%
by various linguists who
work on them categorized

00:02:28.830 --> 00:02:30.150 align:middle line:90%
by various phenomena.

00:02:30.150 --> 00:02:31.860 align:middle line:84%
So there are all
these parameters

00:02:31.860 --> 00:02:32.920 align:middle line:90%
that you can look up.

00:02:32.920 --> 00:02:34.340 align:middle line:84%
You can find out
how many of these

00:02:34.340 --> 00:02:36.090 align:middle line:84%
languages have the
verb before the object.

00:02:36.090 --> 00:02:38.132 align:middle line:84%
How many of them have the
object before the verb.

00:02:38.132 --> 00:02:40.920 align:middle line:84%
You can get it to make you a
map of where these languages are

00:02:40.920 --> 00:02:42.180 align:middle line:90%
and all that cute stuff.

00:02:42.180 --> 00:02:46.260 align:middle line:84%
And they take a stab at
having a typlogically balanced

00:02:46.260 --> 00:02:48.370 align:middle line:84%
genealogically-balanced
survey, I guess.

00:02:48.370 --> 00:02:50.227 align:middle line:84%
So the languages are
not all Indo-European.

00:02:50.227 --> 00:02:51.560 align:middle line:90%
They're from all over the world.

00:02:51.560 --> 00:02:53.885 align:middle line:84%
There's just all
kinds of stuff there.

00:02:53.885 --> 00:02:56.010 align:middle line:84%
Some of the generalizations
that I was showing you,

00:02:56.010 --> 00:03:00.190 align:middle line:84%
it's trivial to show that they
are statistically significant.

00:03:00.190 --> 00:03:03.420 align:middle line:84%
So, for example, there
are according to WALS,

00:03:03.420 --> 00:03:05.700 align:middle line:84%
roughly a third of
the world's languages

00:03:05.700 --> 00:03:07.470 align:middle line:90%
have wh- movement to the left.

00:03:07.470 --> 00:03:09.780 align:middle line:90%
That is, they're like English.

00:03:09.780 --> 00:03:12.630 align:middle line:84%
And the other 2/3
have wh- in situ.

00:03:12.630 --> 00:03:16.320 align:middle line:84%
So if that is wh- goes wherever
the wh- would normally go,

00:03:16.320 --> 00:03:17.880 align:middle line:90%
it doesn't move.

00:03:17.880 --> 00:03:20.640 align:middle line:84%
So when I told you that there
is wh- movement to the left

00:03:20.640 --> 00:03:23.460 align:middle line:84%
and there isn't wh-
movement to the right,

00:03:23.460 --> 00:03:25.650 align:middle line:84%
I'm not a
mathematician, but that

00:03:25.650 --> 00:03:27.750 align:middle line:84%
looks statistically
significant, given

00:03:27.750 --> 00:03:30.810 align:middle line:84%
that they're studying thousands
and thousands of languages.

00:03:30.810 --> 00:03:36.270 align:middle line:84%
It's surprising if you thought
that wh- movement to the right

00:03:36.270 --> 00:03:39.690 align:middle line:84%
was, should be, just as common
as the movement to the left.

00:03:39.690 --> 00:03:42.090 align:middle line:84%
The absence of that in any of
the languages we've studied

00:03:42.090 --> 00:03:44.440 align:middle line:90%
is surprising and significant.

00:03:44.440 --> 00:03:50.370 align:middle line:84%
But as I say, I'll write up
some stats and send them to you.

00:03:50.370 --> 00:03:54.670 align:middle line:84%
Questions about that or anything
else that we did last time?

00:03:54.670 --> 00:03:56.170 align:middle line:90%
OK.

00:03:56.170 --> 00:03:58.180 align:middle line:84%
This, as I said, we had
gotten started on this,

00:03:58.180 --> 00:04:02.140 align:middle line:84%
and I'll just back up and
go through this again.

00:04:02.140 --> 00:04:05.380 align:middle line:84%
We were talking last time
about restrictions on movement.

00:04:05.380 --> 00:04:08.200 align:middle line:84%
So I've given you the
beginnings of some reasons

00:04:08.200 --> 00:04:10.900 align:middle line:84%
to believe that
there are cases where

00:04:10.900 --> 00:04:14.500 align:middle line:84%
forming a tree for a sentence,
forming a syntactic structure,

00:04:14.500 --> 00:04:16.985 align:middle line:84%
isn't just a matter of
merging things to each other,

00:04:16.985 --> 00:04:19.360 align:middle line:84%
but that, sometimes, you take
something which has already

00:04:19.360 --> 00:04:21.740 align:middle line:84%
been merged and you
move it someplace else.

00:04:21.740 --> 00:04:24.220 align:middle line:84%
And we've talked about
different kinds of movement.

00:04:24.220 --> 00:04:26.290 align:middle line:84%
And last time, towards
the end of last time,

00:04:26.290 --> 00:04:29.680 align:middle line:84%
I was starting to show you
examples kinds of movement

00:04:29.680 --> 00:04:31.450 align:middle line:90%
that you couldn't do.

00:04:31.450 --> 00:04:34.360 align:middle line:84%
So I think one of
the examples we

00:04:34.360 --> 00:04:40.000 align:middle line:84%
talked about, you can
say things like "She

00:04:40.000 --> 00:04:49.330 align:middle line:90%
ordered a hamburger and fries."

00:04:49.330 --> 00:04:53.080 align:middle line:84%
But that if you were to take
"fries," replace it with a wh-

00:04:53.080 --> 00:04:57.160 align:middle line:84%
phrase, and tried
to wh- move it,

00:04:57.160 --> 00:05:02.950 align:middle line:84%
giving you "What did she
order a hamburger and?"--

00:05:02.950 --> 00:05:04.630 align:middle line:84%
even if you spell
"did" correctly,

00:05:04.630 --> 00:05:08.110 align:middle line:84%
which is surprisingly
challenging!--

00:05:08.110 --> 00:05:10.070 align:middle line:90%
the result is very bad.

00:05:10.070 --> 00:05:11.920 align:middle line:84%
So I showed you some
examples like that,

00:05:11.920 --> 00:05:15.460 align:middle line:84%
where if you
arrange to put a wh-

00:05:15.460 --> 00:05:17.170 align:middle line:84%
phrase in a certain
kind of structure,

00:05:17.170 --> 00:05:20.240 align:middle line:84%
it somehow can't escape, calling
those structures islands.

00:05:20.240 --> 00:05:21.972 align:middle line:84%
There's some kinds
of wh- movement

00:05:21.972 --> 00:05:22.930 align:middle line:90%
that you just can't do.

00:05:22.930 --> 00:05:24.263 align:middle line:90%
Does this sound familiar at all?

00:05:24.263 --> 00:05:26.450 align:middle line:90%
We were doing this last time.

00:05:26.450 --> 00:05:29.800 align:middle line:84%
And then I said, so
there's big literature

00:05:29.800 --> 00:05:31.690 align:middle line:84%
on syntax identifying
these islands

00:05:31.690 --> 00:05:34.065 align:middle line:84%
and trying to figure out what
are the structures that you

00:05:34.065 --> 00:05:35.200 align:middle line:90%
can't move out of.

00:05:35.200 --> 00:05:37.330 align:middle line:84%
There's actually--
and then attempts

00:05:37.330 --> 00:05:40.150 align:middle line:84%
to develop theories of why
certain things are islands.

00:05:40.150 --> 00:05:43.010 align:middle line:84%
Why are certain kinds of
things opaque to extraction.

00:05:43.010 --> 00:05:47.080 align:middle line:84%
And I wanted to show you one
theory of one particular set

00:05:47.080 --> 00:05:52.180 align:middle line:84%
of examples like this, examples
of places that you cannot move

00:05:52.180 --> 00:05:55.060 align:middle line:90%
out of, because, well, they--

00:05:55.060 --> 00:05:58.810 align:middle line:84%
the explanation for why these
particular things are opaque

00:05:58.810 --> 00:06:02.390 align:middle line:90%
is particularly straightforward.

00:06:02.390 --> 00:06:04.660 align:middle line:84%
So old observation,
actually, that a number

00:06:04.660 --> 00:06:07.390 align:middle line:84%
of the restrictions on
movement that people have found

00:06:07.390 --> 00:06:10.210 align:middle line:84%
can be unified into a
single condition that

00:06:10.210 --> 00:06:13.090 align:middle line:84%
has been called various things
and formalized various ways.

00:06:13.090 --> 00:06:15.010 align:middle line:84%
But today, I'll call
it shortest move.

00:06:15.010 --> 00:06:18.190 align:middle line:84%
It says, basically, if you have
a choice between two movement

00:06:18.190 --> 00:06:21.530 align:middle line:84%
operations, you should
pick the shorter one.

00:06:21.530 --> 00:06:24.460 align:middle line:84%
So pick short moves
over long moves.

00:06:24.460 --> 00:06:29.350 align:middle line:84%
And I gave you this
definition of "short."

00:06:29.350 --> 00:06:31.510 align:middle line:84%
So one of the things people
do in this literature

00:06:31.510 --> 00:06:34.120 align:middle line:84%
is try to figure out exactly
how we ought to define "short."

00:06:34.120 --> 00:06:38.260 align:middle line:84%
So one way to define it would be
to say, one movement is shorter

00:06:38.260 --> 00:06:40.270 align:middle line:84%
than another movement,
if it crosses

00:06:40.270 --> 00:06:41.960 align:middle line:90%
a smaller number of nodes.

00:06:41.960 --> 00:06:45.430 align:middle line:84%
So you start off dominated
by a certain number of nodes,

00:06:45.430 --> 00:06:47.560 align:middle line:84%
and you move to a
position where you are not

00:06:47.560 --> 00:06:50.740 align:middle line:84%
dominated by some of the nodes
that used to dominate you.

00:06:50.740 --> 00:06:52.270 align:middle line:90%
Let's call that your path.

00:06:52.270 --> 00:06:55.300 align:middle line:84%
So the path is the
set of nodes that

00:06:55.300 --> 00:06:58.120 align:middle line:84%
dominate the position the
moved item came from and don't

00:06:58.120 --> 00:07:00.798 align:middle line:84%
dominate the position
where it lands.

00:07:00.798 --> 00:07:02.590 align:middle line:84%
And what we're saying
is you want your path

00:07:02.590 --> 00:07:03.950 align:middle line:90%
to be as short as possible.

00:07:03.950 --> 00:07:06.507 align:middle line:84%
So if you're
comparing two moves,

00:07:06.507 --> 00:07:08.590 align:middle line:84%
and then there are various
ways to formalize this,

00:07:08.590 --> 00:07:10.420 align:middle line:84%
but one is, if you're
comparing to moves,

00:07:10.420 --> 00:07:13.030 align:middle line:84%
count the nodes in the
paths, and pick the one

00:07:13.030 --> 00:07:15.080 align:middle line:90%
that has the smaller number.

00:07:15.080 --> 00:07:16.930 align:middle line:84%
Another way to
formalize it is to say,

00:07:16.930 --> 00:07:19.480 align:middle line:84%
this only works if you're
comparing two nodes that

00:07:19.480 --> 00:07:21.700 align:middle line:84%
have overlapping
paths and you are

00:07:21.700 --> 00:07:24.400 align:middle line:84%
supposed to pick the one that
is a subset, the path that's

00:07:24.400 --> 00:07:26.890 align:middle line:90%
a subset.

00:07:26.890 --> 00:07:29.120 align:middle line:90%
So that's a story to tell here.

00:07:29.120 --> 00:07:30.620 align:middle line:90%
I'll show you some examples.

00:07:30.620 --> 00:07:32.620 align:middle line:84%
The first example, which
I showed you last time,

00:07:32.620 --> 00:07:34.070 align:middle line:84%
was the head
movement constraint.

00:07:34.070 --> 00:07:39.770 align:middle line:84%
So you can move heads but only
over very short distances.

00:07:39.770 --> 00:07:46.570 align:middle line:84%
So English forms yes/no
questions by moving T into C.

00:07:46.570 --> 00:07:50.380 align:middle line:84%
So you can say, ask questions,
like "Will Mary type novels?"

00:07:50.380 --> 00:07:53.543 align:middle line:84%
where you take what's
in T, which is "will,"

00:07:53.543 --> 00:07:55.210 align:middle line:84%
and move it into C.
That's a way to form

00:07:55.210 --> 00:07:56.800 align:middle line:90%
that kind of yes/no question.

00:07:56.800 --> 00:08:00.340 align:middle line:84%
But I said, you
do not, in English

00:08:00.340 --> 00:08:04.420 align:middle line:84%
or in any language, form yes/no
questions by moving the verb

00:08:04.420 --> 00:08:06.580 align:middle line:90%
into C past the auxiliary.

00:08:06.580 --> 00:08:11.410 align:middle line:84%
So English can't ask questions
like "Type Mary will novels?"

00:08:11.410 --> 00:08:13.360 align:middle line:90%
This is out.

00:08:13.360 --> 00:08:15.610 align:middle line:84%
I'm showing you this with
yes/no questions in English.

00:08:15.610 --> 00:08:17.620 align:middle line:84%
I could have shown
it to you with V2.

00:08:17.620 --> 00:08:22.120 align:middle line:84%
So the same thing holds in V2,
verb second, German phenomenon,

00:08:22.120 --> 00:08:28.420 align:middle line:84%
and many other languages in
which something has to be in C.

00:08:28.420 --> 00:08:31.750 align:middle line:84%
And what happens is, whatever
head is highest, moves into C.

00:08:31.750 --> 00:08:33.490 align:middle line:90%
You do the shortest move.

00:08:33.490 --> 00:08:35.230 align:middle line:84%
In German, you
move the auxiliary,

00:08:35.230 --> 00:08:37.450 align:middle line:84%
and if there is no auxiliary,
you move the main verb.

00:08:37.450 --> 00:08:38.867 align:middle line:84%
If there are
multiple auxiliaries,

00:08:38.867 --> 00:08:41.200 align:middle line:90%
you move the highest one.

00:08:41.200 --> 00:08:43.720 align:middle line:84%
So this is one example
of shortest move.

00:08:43.720 --> 00:08:46.300 align:middle line:90%


00:08:46.300 --> 00:08:47.140 align:middle line:90%
Oh, I did it again.

00:08:47.140 --> 00:08:48.670 align:middle line:90%
Sorry, I haven't fixed this yet.

00:08:48.670 --> 00:08:52.180 align:middle line:84%
The path from I to C, which is
really the path from T to C,

00:08:52.180 --> 00:08:55.720 align:middle line:84%
sorry, the path from where
will is to where C is,

00:08:55.720 --> 00:08:57.640 align:middle line:90%
consists of T bar and TP.

00:08:57.640 --> 00:09:00.670 align:middle line:84%
That is, those are the nodes
that dominate T and don't

00:09:00.670 --> 00:09:02.350 align:middle line:90%
dominate C.

00:09:02.350 --> 00:09:05.200 align:middle line:84%
But the path from
V to C consists

00:09:05.200 --> 00:09:08.357 align:middle line:90%
of VP, as well as T bar and TP.

00:09:08.357 --> 00:09:10.690 align:middle line:84%
Do people see that, even
though I wrote the wrong things

00:09:10.690 --> 00:09:11.360 align:middle line:90%
on the slide?

00:09:11.360 --> 00:09:13.780 align:middle line:84%
So please mentally
search, replace I with T,

00:09:13.780 --> 00:09:14.860 align:middle line:90%
on this slide.

00:09:14.860 --> 00:09:17.380 align:middle line:84%
I'm using an old
term for IP, which

00:09:17.380 --> 00:09:19.670 align:middle line:90%
I meant to go back and fix.

00:09:19.670 --> 00:09:22.320 align:middle line:90%
There are no IPs anymore.

00:09:22.320 --> 00:09:23.940 align:middle line:84%
So this is an example
of how you use

00:09:23.940 --> 00:09:27.420 align:middle line:84%
paths to explain
to yourself what

00:09:27.420 --> 00:09:29.950 align:middle line:84%
you mean when you say that one
move is shorter than another.

00:09:29.950 --> 00:09:32.040 align:middle line:84%
As you can see, this is also a
case in which one of the paths

00:09:32.040 --> 00:09:33.082 align:middle line:90%
is a subset of the other.

00:09:33.082 --> 00:09:36.260 align:middle line:90%


00:09:36.260 --> 00:09:37.850 align:middle line:84%
And then the second
kind of case,

00:09:37.850 --> 00:09:40.880 align:middle line:84%
which we talked about last time,
and this is one of the reasons

00:09:40.880 --> 00:09:42.350 align:middle line:84%
that this is only
as far as we got

00:09:42.350 --> 00:09:44.988 align:middle line:84%
because this case is a little
shakier than the first one.

00:09:44.988 --> 00:09:47.030 align:middle line:84%
The head movement constraint
is extremely strong.

00:09:47.030 --> 00:09:48.863 align:middle line:84%
If you violate the head
movement constraint,

00:09:48.863 --> 00:09:51.830 align:middle line:84%
then your sentences are
almost uninterpretable.

00:09:51.830 --> 00:09:54.500 align:middle line:84%
Superiority--
there are all kinds

00:09:54.500 --> 00:09:56.090 align:middle line:84%
of complicated
semantic things going

00:09:56.090 --> 00:10:00.590 align:middle line:84%
on that allow you to
sometimes violate superiority.

00:10:00.590 --> 00:10:03.590 align:middle line:84%
But, at least in some
cases, out of the blue, if I

00:10:03.590 --> 00:10:05.840 align:middle line:90%
ask you "Who bought what?"

00:10:05.840 --> 00:10:08.090 align:middle line:84%
that's better than asking
"What did who buy?"

00:10:08.090 --> 00:10:10.940 align:middle line:84%
And I was getting general
agreement with that last time.

00:10:10.940 --> 00:10:14.630 align:middle line:84%
One of the things I said last
time, and I think this is true,

00:10:14.630 --> 00:10:21.020 align:middle line:84%
is that superiority judgments
are particularly strong if you

00:10:21.020 --> 00:10:22.010 align:middle line:90%
are talking about--

00:10:22.010 --> 00:10:24.530 align:middle line:90%


00:10:24.530 --> 00:10:26.450 align:middle line:90%
if there are only two answers.

00:10:26.450 --> 00:10:29.630 align:middle line:84%
So wh- questions--
multiple wh- questions,

00:10:29.630 --> 00:10:34.040 align:middle line:84%
what you're asking for is
a list of pairs, usually.

00:10:34.040 --> 00:10:35.780 align:middle line:84%
So if I ask you
"Who bought what?"

00:10:35.780 --> 00:10:38.450 align:middle line:84%
the answer is going
to be, well, you know,

00:10:38.450 --> 00:10:41.570 align:middle line:84%
"Sally bought chocolate,
and Bill novels,

00:10:41.570 --> 00:10:44.210 align:middle line:90%
and Mary bought a computer."

00:10:44.210 --> 00:10:46.730 align:middle line:84%
So it'll be a list
of people and things,

00:10:46.730 --> 00:10:48.863 align:middle line:84%
such that the people
bought the things.

00:10:48.863 --> 00:10:50.780 align:middle line:84%
Superiority-- we talked
about this last time--

00:10:50.780 --> 00:10:52.580 align:middle line:84%
but superiority
judgments are strongest

00:10:52.580 --> 00:10:54.660 align:middle line:90%
when there's only one pair.

00:10:54.660 --> 00:11:00.980 align:middle line:84%
So if you see two
people fighting,

00:11:00.980 --> 00:11:05.750 align:middle line:84%
you might ask "Who
hit who first?"

00:11:05.750 --> 00:11:08.480 align:middle line:90%


00:11:08.480 --> 00:11:13.925 align:middle line:84%
Compare that with "Who
did who hit first?"

00:11:13.925 --> 00:11:16.760 align:middle line:90%


00:11:16.760 --> 00:11:19.580 align:middle line:84%
For me, at least, that's a
particularly strong judgment.

00:11:19.580 --> 00:11:21.530 align:middle line:84%
Are there people
who prefer one--

00:11:21.530 --> 00:11:22.040 align:middle line:90%
let's see.

00:11:22.040 --> 00:11:24.740 align:middle line:84%
Let's see a show of hands,
who prefers the first of these

00:11:24.740 --> 00:11:26.570 align:middle line:90%
to the second?

00:11:26.570 --> 00:11:28.940 align:middle line:84%
Who prefers the
second to the first?

00:11:28.940 --> 00:11:31.520 align:middle line:84%
Who finds them both fine and
is not sure why some of you

00:11:31.520 --> 00:11:32.960 align:middle line:84%
raised your hands
the first time?

00:11:32.960 --> 00:11:36.950 align:middle line:84%
So there are a few of you
who don't care whether there

00:11:36.950 --> 00:11:38.090 align:middle line:90%
is only a single pair.

00:11:38.090 --> 00:11:41.870 align:middle line:84%
But for many of you, this
one is better than this one.

00:11:41.870 --> 00:11:45.390 align:middle line:84%
There's something
wrong with this.

00:11:45.390 --> 00:11:47.965 align:middle line:84%
So this is a phenomenon which
is not as strong as the head

00:11:47.965 --> 00:11:50.090 align:middle line:84%
movement constraint but
resembles the head movement

00:11:50.090 --> 00:11:53.180 align:middle line:84%
constraint, in that,
there's a preference when

00:11:53.180 --> 00:11:56.027 align:middle line:84%
you've got a choice between two
things which of them to move.

00:11:56.027 --> 00:11:58.610 align:middle line:84%
There's a preference for moving
the higher one, the first one.

00:11:58.610 --> 00:12:01.637 align:middle line:90%


00:12:01.637 --> 00:12:03.470 align:middle line:84%
And then here's the
case number three, which

00:12:03.470 --> 00:12:05.900 align:middle line:90%
we didn't get to last time.

00:12:05.900 --> 00:12:08.450 align:middle line:84%
These are what are
called wh- islands.

00:12:08.450 --> 00:12:11.330 align:middle line:84%
"Island," again, is
a name syntacticians

00:12:11.330 --> 00:12:13.700 align:middle line:84%
use for these positions
that you can't move out of.

00:12:13.700 --> 00:12:16.640 align:middle line:84%
There's this metaphor, I guess,
that involves wh- phrases

00:12:16.640 --> 00:12:18.230 align:middle line:90%
not being able to swim.

00:12:18.230 --> 00:12:21.650 align:middle line:84%
If they're stuck on an
island, then they're stuck.

00:12:21.650 --> 00:12:27.140 align:middle line:84%
So here's a sentence with
an embedded wh- question.

00:12:27.140 --> 00:12:30.080 align:middle line:84%
"I wonder what
Bill gave to Fred."

00:12:30.080 --> 00:12:33.410 align:middle line:84%
So this whole sentence
is not a question really.

00:12:33.410 --> 00:12:34.250 align:middle line:90%
It's a statement.

00:12:34.250 --> 00:12:36.770 align:middle line:84%
I'm telling you something
about my mental state.

00:12:36.770 --> 00:12:39.650 align:middle line:84%
My mental state is one in
which I don't know the answer

00:12:39.650 --> 00:12:41.700 align:middle line:90%
to a particular question.

00:12:41.700 --> 00:12:44.390 align:middle line:84%
And that question is, "What
did Bill give to Fred?"

00:12:44.390 --> 00:12:46.400 align:middle line:84%
So there's wh-
movements happening

00:12:46.400 --> 00:12:47.540 align:middle line:90%
in the embedded clause.

00:12:47.540 --> 00:12:49.880 align:middle line:84%
What has moved to the edge
of the embedded clause.

00:12:49.880 --> 00:12:51.110 align:middle line:90%
Is that clear?

00:12:51.110 --> 00:12:52.670 align:middle line:90%
Does that make sense?

00:12:52.670 --> 00:12:54.960 align:middle line:84%
That's what's going
on in this sentence.

00:12:54.960 --> 00:12:58.990 align:middle line:84%
Now observation about
embedded questions.

00:12:58.990 --> 00:13:01.300 align:middle line:84%
If I ask you, "Who do you
wonder what Bill gave to?"

00:13:01.300 --> 00:13:04.080 align:middle line:90%


00:13:04.080 --> 00:13:07.980 align:middle line:90%
you are upset.

00:13:07.980 --> 00:13:09.580 align:middle line:90%
So what did I do.

00:13:09.580 --> 00:13:15.750 align:middle line:84%
Well, I turned Fred into a
wh- question, a wh- word, zap.

00:13:15.750 --> 00:13:19.590 align:middle line:84%
Now Fred is who,
and I moved "who"

00:13:19.590 --> 00:13:22.980 align:middle line:84%
out of the embedded clause
into the matrix clause.

00:13:22.980 --> 00:13:24.610 align:middle line:84%
"Who do you wonder
what Bill gave to?"

00:13:24.610 --> 00:13:26.700 align:middle line:84%
where the answer is,
well, I'm wondering

00:13:26.700 --> 00:13:29.085 align:middle line:90%
what Bill gave to Fred.

00:13:29.085 --> 00:13:31.450 align:middle line:84%
It ought to be able to
mean that, but it really,

00:13:31.450 --> 00:13:32.830 align:middle line:90%
really can't.

00:13:32.830 --> 00:13:37.230 align:middle line:84%
There's something badly
wrong with this sentence.

00:13:37.230 --> 00:13:39.030 align:middle line:90%
So this is called a wh- island.

00:13:39.030 --> 00:13:42.690 align:middle line:84%
A wh- question, like "what Bill
gave to Fred," is an island.

00:13:42.690 --> 00:13:43.440 align:middle line:90%
You can form that.

00:13:43.440 --> 00:13:45.220 align:middle line:84%
That's what the
first sentence says.

00:13:45.220 --> 00:13:46.637 align:middle line:84%
"I wonder what
Bill gave to Fred."

00:13:46.637 --> 00:13:49.900 align:middle line:84%
But you can't wh-
extract out of it.

00:13:49.900 --> 00:13:52.470 align:middle line:90%
The proposal people have made--

00:13:52.470 --> 00:13:54.720 align:middle line:84%
well, look, there are at
least two possible things

00:13:54.720 --> 00:13:56.310 align:middle line:84%
that could be wrong with
the green arrow there,

00:13:56.310 --> 00:13:58.185 align:middle line:84%
the one that's going
from the embedded clause

00:13:58.185 --> 00:13:59.700 align:middle line:90%
into the matrix clause.

00:13:59.700 --> 00:14:03.540 align:middle line:84%
One is, here I am
in the matrix CP,

00:14:03.540 --> 00:14:06.040 align:middle line:84%
and I'm trying to
decide what to move.

00:14:06.040 --> 00:14:08.400 align:middle line:90%
And there are two wh- phrases.

00:14:08.400 --> 00:14:09.840 align:middle line:84%
There's "what"
and there's "who."

00:14:09.840 --> 00:14:12.160 align:middle line:84%
I should really
have a tree here.

00:14:12.160 --> 00:14:15.315 align:middle line:90%
Let me draw a quasi tree.

00:14:15.315 --> 00:14:24.442 align:middle line:90%


00:14:24.442 --> 00:14:25.025 align:middle line:90%
To "wonder"...

00:14:25.025 --> 00:14:30.550 align:middle line:90%


00:14:30.550 --> 00:14:33.450 align:middle line:90%
Here we have "what."

00:14:33.450 --> 00:14:36.790 align:middle line:84%
Then, I'm out of words, so
I'll just use a triangle here,

00:14:36.790 --> 00:14:39.580 align:middle line:90%
to who.

00:14:39.580 --> 00:14:43.893 align:middle line:84%
So this is "You wonder
what John gave to who."

00:14:43.893 --> 00:14:45.310 align:middle line:84%
I know my handwriting
is terrible,

00:14:45.310 --> 00:14:47.920 align:middle line:84%
but can people see that that's
what that tree is meant to be?

00:14:47.920 --> 00:14:50.818 align:middle line:84%
So this is the
tree for what you--

00:14:50.818 --> 00:14:53.110 align:middle line:84%
for the second sentence, this
is going to be my attempt

00:14:53.110 --> 00:14:54.650 align:middle line:90%
to draw the second sentence.

00:14:54.650 --> 00:14:57.610 align:middle line:84%
So "Who do you wonder
what Bill gave to?"

00:14:57.610 --> 00:15:01.120 align:middle line:84%
So we've got-- we can fill
in some of the gaps here.

00:15:01.120 --> 00:15:05.320 align:middle line:90%
"Bill gave (blank) to who."

00:15:05.320 --> 00:15:08.470 align:middle line:84%
So we first moved "what" into
the specifier of the embedded

00:15:08.470 --> 00:15:10.090 align:middle line:90%
CP.

00:15:10.090 --> 00:15:14.380 align:middle line:84%
And now, this is one of
the few times in my life

00:15:14.380 --> 00:15:16.840 align:middle line:90%
I wish I had colored chalk.

00:15:16.840 --> 00:15:18.640 align:middle line:84%
Now we're going to
"who," and we're

00:15:18.640 --> 00:15:21.940 align:middle line:90%
going to move it up to here.

00:15:21.940 --> 00:15:23.980 align:middle line:84%
And there's something
wrong with this.

00:15:23.980 --> 00:15:25.887 align:middle line:84%
Try to figure out
what's wrong with it.

00:15:25.887 --> 00:15:27.970 align:middle line:84%
The shortest move story
about what's wrong with it

00:15:27.970 --> 00:15:28.610 align:middle line:90%
goes like this.

00:15:28.610 --> 00:15:29.860 align:middle line:84%
Well, actually,
there are two things

00:15:29.860 --> 00:15:31.068 align:middle line:90%
that you could blame this on.

00:15:31.068 --> 00:15:32.920 align:middle line:84%
One is, when we're
here and we're

00:15:32.920 --> 00:15:35.290 align:middle line:84%
trying to decide what to
move into this position,

00:15:35.290 --> 00:15:37.750 align:middle line:84%
well, at this point,
there are two wh- phrases,

00:15:37.750 --> 00:15:38.740 align:middle line:90%
"what" and "who."

00:15:38.740 --> 00:15:41.590 align:middle line:90%
And "what" is higher.

00:15:41.590 --> 00:15:44.170 align:middle line:84%
So maybe one thing
that's wrong with it

00:15:44.170 --> 00:15:46.955 align:middle line:84%
is, basically,
superiority over again.

00:15:46.955 --> 00:15:48.310 align:middle line:90%
You've got two wh- phrases.

00:15:48.310 --> 00:15:49.768 align:middle line:84%
You have to pick
which one to move.

00:15:49.768 --> 00:15:52.620 align:middle line:84%
You should move the
higher one, if you didn't.

00:15:52.620 --> 00:15:56.453 align:middle line:84%
So that you didn't
pick the shorter move.

00:15:56.453 --> 00:15:58.870 align:middle line:84%
And the other objection you
might have to this green arrow

00:15:58.870 --> 00:16:00.520 align:middle line:90%
goes like this.

00:16:00.520 --> 00:16:02.470 align:middle line:84%
The wh- movement is
to the spec of CP.

00:16:02.470 --> 00:16:04.120 align:middle line:90%
Here's "who" sitting here.

00:16:04.120 --> 00:16:06.520 align:middle line:84%
If you're going to
move "who," well,

00:16:06.520 --> 00:16:07.990 align:middle line:90%
you shouldn't move it here.

00:16:07.990 --> 00:16:09.995 align:middle line:90%
You should move it here.

00:16:09.995 --> 00:16:10.495 align:middle line:90%
Sorry.

00:16:10.495 --> 00:16:12.400 align:middle line:84%
Let me do that without
putting my body in the way.

00:16:12.400 --> 00:16:13.525 align:middle line:90%
You shouldn't move it here.

00:16:13.525 --> 00:16:15.820 align:middle line:84%
Here's a spec of CP to
which it's going to land.

00:16:15.820 --> 00:16:17.740 align:middle line:84%
But really what you
should do is move it here.

00:16:17.740 --> 00:16:20.620 align:middle line:84%
And you can't, because
there's a wh- phrase here.

00:16:20.620 --> 00:16:22.120 align:middle line:90%
This phrase is stuck.

00:16:22.120 --> 00:16:27.370 align:middle line:84%
So the ill-formedness of
the sentence, the star,

00:16:27.370 --> 00:16:32.350 align:middle line:84%
is this feeling
you're having trying

00:16:32.350 --> 00:16:36.370 align:middle line:84%
to satisfy multiple
conflicting requirements.

00:16:36.370 --> 00:16:38.300 align:middle line:84%
There's a Greek
tragedy going on here.

00:16:38.300 --> 00:16:40.330 align:middle line:84%
"Who" absolutely
needs to land here,

00:16:40.330 --> 00:16:41.720 align:middle line:90%
but there's "what" in the way.

00:16:41.720 --> 00:16:45.100 align:middle line:84%
And so, "who" wants to kill its
father and marry its mother,

00:16:45.100 --> 00:16:47.410 align:middle line:90%
and then, bad things happen.

00:16:47.410 --> 00:16:49.798 align:middle line:84%
You can't do everything
that you want to do.

00:16:49.798 --> 00:16:52.090 align:middle line:84%
So this wants to make the
shortest possible move and it

00:16:52.090 --> 00:16:52.690 align:middle line:90%
can't.

00:16:52.690 --> 00:16:54.460 align:middle line:90%
And it's unhappy and it pouts.

00:16:54.460 --> 00:16:56.950 align:middle line:84%
And that's got the
shape of a star.

00:16:56.950 --> 00:17:00.730 align:middle line:84%
That's a story, anyway, about
why wh- islands are islands.

00:17:00.730 --> 00:17:02.686 align:middle line:90%
Do people see the problem?

00:17:02.686 --> 00:17:04.160 align:middle line:90%
That, that sentence is bad.

00:17:04.160 --> 00:17:06.310 align:middle line:84%
And it's not like it's
unclear what it means,

00:17:06.310 --> 00:17:09.170 align:middle line:90%
but you just can't say that.

00:17:09.170 --> 00:17:09.898 align:middle line:90%
Yeah?

00:17:09.898 --> 00:17:11.190 align:middle line:90%
AUDIENCE: How would you say it?

00:17:11.190 --> 00:17:13.319 align:middle line:90%
NORVIN RICHARDS: You can't.

00:17:13.319 --> 00:17:15.329 align:middle line:84%
"Who do you wonder
what Bill gave to?"

00:17:15.329 --> 00:17:17.579 align:middle line:84%
The closest you can come
to saying it in English

00:17:17.579 --> 00:17:19.560 align:middle line:84%
is to say things
like, "Who is it

00:17:19.560 --> 00:17:24.609 align:middle line:84%
that you're wondering what
Bill gave to him or her."

00:17:24.609 --> 00:17:28.140 align:middle line:84%
So sometimes, for some
people, putting in a pronoun

00:17:28.140 --> 00:17:30.180 align:middle line:90%
improves these sentences.

00:17:30.180 --> 00:17:32.940 align:middle line:84%
For me, the effect
is not so strong.

00:17:32.940 --> 00:17:35.250 align:middle line:84%
There isn't a very
good way to say this.

00:17:35.250 --> 00:17:37.320 align:middle line:84%
There are languages
in which this is OK,

00:17:37.320 --> 00:17:41.010 align:middle line:84%
which is an interesting--
there are questions about why.

00:17:41.010 --> 00:17:42.870 align:middle line:90%
But in English, not so much.

00:17:42.870 --> 00:17:43.920 align:middle line:90%
Yes?

00:17:43.920 --> 00:17:47.690 align:middle line:84%
AUDIENCE: What about "What do
you wonder Bill gave to who?"

00:17:47.690 --> 00:17:50.630 align:middle line:84%
NORVIN RICHARDS: "What do
you wonder Bill gave to who?"

00:17:50.630 --> 00:18:05.170 align:middle line:84%
So the point-- so one of
the things I just said was,

00:18:05.170 --> 00:18:08.292 align:middle line:84%
when we're asking what went
wrong with the sentence that

00:18:08.292 --> 00:18:10.750 align:middle line:84%
I've got up there, "Who do you
wonder what Bill gave to?"--

00:18:10.750 --> 00:18:13.480 align:middle line:90%


00:18:13.480 --> 00:18:16.630 align:middle line:84%
one thing that you could blame
it on is, you moved "who,"

00:18:16.630 --> 00:18:17.950 align:middle line:90%
and "what" is closer.

00:18:17.950 --> 00:18:21.850 align:middle line:84%
What you should really
have done is moved "what."

00:18:21.850 --> 00:18:23.770 align:middle line:84%
I don't know about
you guys, but, for me,

00:18:23.770 --> 00:18:26.720 align:middle line:84%
"What do you wonder
Bill gave to who?"

00:18:26.720 --> 00:18:29.620 align:middle line:90%
I can't say.

00:18:29.620 --> 00:18:32.590 align:middle line:84%
So this is another one of
these Greek tragedy things.

00:18:32.590 --> 00:18:36.550 align:middle line:84%
Shortest move really wants this
to be the thing that moves.

00:18:36.550 --> 00:18:38.830 align:middle line:84%
But the result is
that, let's think

00:18:38.830 --> 00:18:40.120 align:middle line:90%
about what wh- movement does.

00:18:40.120 --> 00:18:41.770 align:middle line:90%
It makes questions, right?

00:18:41.770 --> 00:18:43.570 align:middle line:84%
So when you move
"what" here, you're

00:18:43.570 --> 00:18:45.130 align:middle line:90%
making this into a question.

00:18:45.130 --> 00:18:48.940 align:middle line:84%
You wonder what Bill gave to
Bill-- to Fred or whatever.

00:18:48.940 --> 00:18:51.400 align:middle line:84%
If you move it up here, you're
making this into a question.

00:18:51.400 --> 00:18:55.000 align:middle line:84%
And maybe what we're seeing
is that you can't do both.

00:18:55.000 --> 00:18:58.440 align:middle line:84%
And we could demonstrate that
without doing an wh- island.

00:18:58.440 --> 00:19:00.210 align:middle line:84%
So it's also bad,
at least for me.

00:19:00.210 --> 00:19:01.470 align:middle line:90%
I'm going to star this.

00:19:01.470 --> 00:19:03.345 align:middle line:84%
It's also bad, at least
for me, to say things

00:19:03.345 --> 00:19:10.325 align:middle line:84%
like what do you wonder
Bill gave to Fred.

00:19:10.325 --> 00:19:12.950 align:middle line:90%


00:19:12.950 --> 00:19:14.480 align:middle line:90%
This is also out.

00:19:14.480 --> 00:19:16.670 align:middle line:84%
So it's just independently
impossible for "what"

00:19:16.670 --> 00:19:19.880 align:middle line:84%
to start here, move
here, and then move here.

00:19:19.880 --> 00:19:24.260 align:middle line:90%
It's as though "what" can't be--

00:19:24.260 --> 00:19:27.020 align:middle line:84%
you're attempting to form two
questions with a single wh-

00:19:27.020 --> 00:19:29.891 align:middle line:90%
word, and you can't.

00:19:29.891 --> 00:19:32.037 align:middle line:84%
AUDIENCE: I thought
the last one is OK.

00:19:32.037 --> 00:19:34.370 align:middle line:84%
NORVIN RICHARDS: "What do you
wonder Bill gave to Fred?"

00:19:34.370 --> 00:19:36.037 align:middle line:84%
AUDIENCE: No. you
could say "What do you

00:19:36.037 --> 00:19:40.270 align:middle line:84%
wonder whether Bill gave to
Fred" is more normal to me,

00:19:40.270 --> 00:19:41.980 align:middle line:90%
but also [INAUDIBLE].

00:19:41.980 --> 00:19:43.030 align:middle line:90%
NORVIN RICHARDS: Yeah.

00:19:43.030 --> 00:19:44.958 align:middle line:84%
For me, at least,
I can't say this.

00:19:44.958 --> 00:19:47.500 align:middle line:84%
But I agree with you, that if
I were to make it, "What do you

00:19:47.500 --> 00:19:49.450 align:middle line:90%
wonder if Bill gave to Fred"--

00:19:49.450 --> 00:19:51.910 align:middle line:84%
that's still bad, but
it's better than this.

00:19:51.910 --> 00:19:53.910 align:middle line:90%
That's my feeling about it.

00:19:53.910 --> 00:19:54.410 align:middle line:90%
Joseph?

00:19:54.410 --> 00:19:57.740 align:middle line:90%


00:19:57.740 --> 00:20:01.080 align:middle line:84%
AUDIENCE: "Wonder" replaced with
"think" seems perfectly fine.

00:20:01.080 --> 00:20:04.250 align:middle line:90%
So [INAUDIBLE].

00:20:04.250 --> 00:20:07.160 align:middle line:90%
NORVIN RICHARDS: So notice--

00:20:07.160 --> 00:20:07.820 align:middle line:90%
good point.

00:20:07.820 --> 00:20:11.200 align:middle line:90%


00:20:11.200 --> 00:20:13.490 align:middle line:90%
AUDIENCE: [INAUDIBLE]

00:20:13.490 --> 00:20:15.080 align:middle line:84%
NORVIN RICHARDS:
Right, that's it.

00:20:15.080 --> 00:20:18.770 align:middle line:84%
So "wonder" selects for
an interrogative CP.

00:20:18.770 --> 00:20:22.430 align:middle line:90%
You can say "I wonder who left?"

00:20:22.430 --> 00:20:30.720 align:middle line:84%
"why they left," "where
they went," but not--

00:20:30.720 --> 00:20:34.740 align:middle line:90%
or "whether they left."

00:20:34.740 --> 00:20:37.890 align:middle line:84%
There are various things you can
say, but they're all questions.

00:20:37.890 --> 00:20:39.930 align:middle line:84%
So "wonder" selects
for a question.

00:20:39.930 --> 00:20:45.150 align:middle line:90%
And "that she left" is no good.

00:20:45.150 --> 00:20:49.890 align:middle line:90%
"Think" is the opposite.

00:20:49.890 --> 00:20:55.830 align:middle line:84%
You can say, "I think that
she left," but not, "I

00:20:55.830 --> 00:21:00.132 align:middle line:84%
think whether she left,"
or "I think we left."

00:21:00.132 --> 00:21:00.840 align:middle line:90%
Things like that.

00:21:00.840 --> 00:21:02.680 align:middle line:84%
Yeah, so you're
absolutely right.

00:21:02.680 --> 00:21:05.710 align:middle line:84%
If I make that "think," "What
do you think Bill gave to Fred?"

00:21:05.710 --> 00:21:07.710 align:middle line:84%
then "what" is going to
be able to move up here,

00:21:07.710 --> 00:21:09.900 align:middle line:84%
because "think"
doesn't care, in fact,

00:21:09.900 --> 00:21:11.490 align:middle line:84%
doesn't want this
to be a question.

00:21:11.490 --> 00:21:13.560 align:middle line:84%
But "wonder" does want
this to be a question,

00:21:13.560 --> 00:21:15.480 align:middle line:84%
with the consequence
that if "what" goes here,

00:21:15.480 --> 00:21:17.040 align:middle line:90%
it's as though it's stuck.

00:21:17.040 --> 00:21:18.030 align:middle line:90%
It's got a job.

00:21:18.030 --> 00:21:19.620 align:middle line:84%
It needs to make
this a question.

00:21:19.620 --> 00:21:22.118 align:middle line:84%
If it moves, then it
isn't a question anymore.

00:21:22.118 --> 00:21:23.160 align:middle line:90%
Something has gone wrong.

00:21:23.160 --> 00:21:25.710 align:middle line:90%


00:21:25.710 --> 00:21:28.010 align:middle line:84%
So we have two
conflicting things.

00:21:28.010 --> 00:21:30.960 align:middle line:84%
Your shortest move that
says, in this example,

00:21:30.960 --> 00:21:33.380 align:middle line:84%
"what" should move up to
here should really be "what."

00:21:33.380 --> 00:21:35.000 align:middle line:84%
And then, you have
"wonder," which says, "No,

00:21:35.000 --> 00:21:36.250 align:middle line:90%
'what,' you have to stay here.

00:21:36.250 --> 00:21:38.040 align:middle line:90%
You have a job."

00:21:38.040 --> 00:21:41.595 align:middle line:84%
And these two things
can't be reconciled.

00:21:41.595 --> 00:21:43.970 align:middle line:84%
Notice that we are out of the
domain of optimality theory

00:21:43.970 --> 00:21:44.678 align:middle line:90%
here.

00:21:44.678 --> 00:21:46.220 align:middle line:84%
So if you remember
optimality theory,

00:21:46.220 --> 00:21:47.480 align:middle line:84%
back when we were
doing phonology,

00:21:47.480 --> 00:21:49.070 align:middle line:84%
we were talking as
though, sometimes,

00:21:49.070 --> 00:21:50.987 align:middle line:84%
when you have two things
and they're fighting,

00:21:50.987 --> 00:21:53.180 align:middle line:84%
you just try to find
some middle ground.

00:21:53.180 --> 00:21:54.860 align:middle line:84%
You find the thing
that's best or you

00:21:54.860 --> 00:21:56.510 align:middle line:90%
decide who's more important.

00:21:56.510 --> 00:21:59.210 align:middle line:84%
That's not the way I usually
settle fights between people,

00:21:59.210 --> 00:22:01.610 align:middle line:84%
but apparently, in optimality
theory, that's what you do.

00:22:01.610 --> 00:22:03.320 align:middle line:84%
You decide which of
these requirements

00:22:03.320 --> 00:22:05.060 align:middle line:90%
is the most important one.

00:22:05.060 --> 00:22:07.010 align:middle line:84%
The way I'm talking
here is not that.

00:22:07.010 --> 00:22:10.460 align:middle line:84%
This is more like, here you
have two conflicting constraints

00:22:10.460 --> 00:22:11.660 align:middle line:90%
and so you die.

00:22:11.660 --> 00:22:13.125 align:middle line:90%
You can't do anything.

00:22:13.125 --> 00:22:14.750 align:middle line:84%
There's no way to
make everybody happy.

00:22:14.750 --> 00:22:19.070 align:middle line:90%


00:22:19.070 --> 00:22:21.800 align:middle line:84%
So this has been an
exercise in, well, first,

00:22:21.800 --> 00:22:24.530 align:middle line:84%
showing you some other places
where movement is barred,

00:22:24.530 --> 00:22:28.435 align:middle line:84%
and then, also,
giving you just a hint

00:22:28.435 --> 00:22:30.560 align:middle line:84%
of a very large and
interesting literature in which

00:22:30.560 --> 00:22:34.070 align:middle line:84%
syntacticians look at
a variety of phenomena

00:22:34.070 --> 00:22:36.020 align:middle line:90%
and try to unify them.

00:22:36.020 --> 00:22:40.087 align:middle line:84%
So head movement constraint,
and superiority, and wh- islands

00:22:40.087 --> 00:22:42.170 align:middle line:84%
are all things that were
discovered independently,

00:22:42.170 --> 00:22:43.760 align:middle line:84%
at different times,
by different people.

00:22:43.760 --> 00:22:45.650 align:middle line:84%
And then another
generation of syntacticians

00:22:45.650 --> 00:22:48.590 align:middle line:84%
came along and said, no, these
all belong in the same box.

00:22:48.590 --> 00:22:50.990 align:middle line:84%
And that's where we
are at this point.

00:22:50.990 --> 00:22:54.140 align:middle line:90%


00:22:54.140 --> 00:22:55.940 align:middle line:84%
Lots more work like
this to talk about.

00:22:55.940 --> 00:22:58.428 align:middle line:84%
But this is 24.900, and
although I am a syntactician--

00:22:58.428 --> 00:22:59.970 align:middle line:84%
I looked at the
syllabus, by the way,

00:22:59.970 --> 00:23:01.762 align:middle line:84%
and we're actually,
more or less, on track.

00:23:01.762 --> 00:23:02.540 align:middle line:90%
I was astonished.

00:23:02.540 --> 00:23:04.813 align:middle line:84%
We're supposed to stop
talking about syntax today,

00:23:04.813 --> 00:23:05.480 align:middle line:90%
and so, we will.

00:23:05.480 --> 00:23:08.220 align:middle line:90%


00:23:08.220 --> 00:23:10.440 align:middle line:84%
One more thing for me to
show you about syntax.

00:23:10.440 --> 00:23:13.770 align:middle line:84%
It's actually kind of connected
to this, to Joseph's point.

00:23:13.770 --> 00:23:15.250 align:middle line:90%
Here's a sentence.

00:23:15.250 --> 00:23:18.700 align:middle line:84%
"She will say that we
should buy kumquats."

00:23:18.700 --> 00:23:19.950 align:middle line:90%
Questions about this sentence?

00:23:19.950 --> 00:23:21.360 align:middle line:90%
Does this make sense?

00:23:21.360 --> 00:23:22.890 align:middle line:84%
Did I put anything
in here that's

00:23:22.890 --> 00:23:26.935 align:middle line:90%
surprising or disturbing?

00:23:26.935 --> 00:23:29.060 align:middle line:84%
I think I have a new thing
in T, which is "should,"

00:23:29.060 --> 00:23:32.240 align:middle line:84%
but maybe that's
not so alarming.

00:23:32.240 --> 00:23:35.690 align:middle line:84%
And now what happens if we
change "kumquats" into a wh-

00:23:35.690 --> 00:23:38.893 align:middle line:84%
word, stop, and we wh-
move it to the beginning

00:23:38.893 --> 00:23:39.560 align:middle line:90%
of the sentence.

00:23:39.560 --> 00:23:44.120 align:middle line:84%
Well, so that's me turning
"kumquats" into "what,

00:23:44.120 --> 00:23:45.980 align:middle line:84%
and also turning the
TP into a triangle,

00:23:45.980 --> 00:23:47.490 align:middle line:84%
because I was
running out of room.

00:23:47.490 --> 00:23:51.620 align:middle line:84%
And then we're going to move
what into the matrix clause,

00:23:51.620 --> 00:23:53.940 align:middle line:84%
so we'll get what will she
say that we should buy.

00:23:53.940 --> 00:23:55.640 align:middle line:84%
So we've done wh-
movement of "what,"

00:23:55.640 --> 00:23:57.890 align:middle line:84%
and we've also done
head movement of "will"

00:23:57.890 --> 00:24:01.500 align:middle line:84%
from T into C. I haven't
drawn an arrow for that.

00:24:01.500 --> 00:24:05.110 align:middle line:84%
And that's how we form
that wh- question.

00:24:05.110 --> 00:24:09.300 align:middle line:84%
But given everything I just
said about shortest move,

00:24:09.300 --> 00:24:10.810 align:middle line:90%
you might worry.

00:24:10.810 --> 00:24:13.610 align:middle line:84%
You might say to
yourself, but wait,

00:24:13.610 --> 00:24:15.770 align:middle line:84%
isn't CT a closer
place to land--

00:24:15.770 --> 00:24:17.413 align:middle line:84%
that lower CT,
that we should buy,

00:24:17.413 --> 00:24:18.830 align:middle line:84%
isn't that a closer
place to land?

00:24:18.830 --> 00:24:23.660 align:middle line:84%
After all, wh- words move
to the specifier of CP.

00:24:23.660 --> 00:24:25.970 align:middle line:84%
Here is an example of
where there are two CPs.

00:24:25.970 --> 00:24:29.090 align:middle line:90%
So what should we do?

00:24:29.090 --> 00:24:32.210 align:middle line:84%
And there are various theories
you can have about that.

00:24:32.210 --> 00:24:34.010 align:middle line:84%
One would be to say,
yeah, a wh- phrase has

00:24:34.010 --> 00:24:36.350 align:middle line:84%
moved to specifer of CP,
but it's specifically

00:24:36.350 --> 00:24:37.370 align:middle line:90%
interrogative CP.

00:24:37.370 --> 00:24:38.780 align:middle line:90%
They form questions.

00:24:38.780 --> 00:24:41.650 align:middle line:90%
So not that kind of CP.

00:24:41.650 --> 00:24:45.280 align:middle line:84%
But another kind of
theory we could have,

00:24:45.280 --> 00:24:48.880 align:middle line:84%
would be to say, no,
shortest move is correct.

00:24:48.880 --> 00:24:51.250 align:middle line:84%
The fact that "what"
can, in principle,

00:24:51.250 --> 00:24:54.730 align:middle line:84%
land in the specifer of
CP, means that it must.

00:24:54.730 --> 00:24:57.670 align:middle line:84%
So it's not possible to do the
arrow-- bless you-- the longer

00:24:57.670 --> 00:24:59.620 align:middle line:84%
arrow that I showed you
on the previous slide,

00:24:59.620 --> 00:25:01.540 align:middle line:84%
where "what" just starts
down there and moves

00:25:01.540 --> 00:25:03.490 align:middle line:90%
in one hop up to there.

00:25:03.490 --> 00:25:07.870 align:middle line:84%
Movement is always-- the name
for this is successive cyclic.

00:25:07.870 --> 00:25:10.210 align:middle line:90%
It's always in little hops.

00:25:10.210 --> 00:25:14.980 align:middle line:84%
You can't move from one CP, out
of that CP, into another CP,

00:25:14.980 --> 00:25:19.240 align:middle line:84%
without stopping at the edge of
the CP that you started out in.

00:25:19.240 --> 00:25:22.990 align:middle line:84%
This is something that was
proposed by Noam Chomsky, who,

00:25:22.990 --> 00:25:23.890 align:middle line:90%
decades ago--

00:25:23.890 --> 00:25:25.960 align:middle line:84%
and when he first
proposed it, it

00:25:25.960 --> 00:25:28.515 align:middle line:84%
was met with
widespread derision.

00:25:28.515 --> 00:25:30.640 align:middle line:84%
People were like "Just
because of the shortest move

00:25:30.640 --> 00:25:31.750 align:middle line:84%
constraint, you're
going to say that we're

00:25:31.750 --> 00:25:33.220 align:middle line:90%
doing all these little hops."

00:25:33.220 --> 00:25:37.390 align:middle line:84%
And then, there was immediately
a flood of really good evidence

00:25:37.390 --> 00:25:39.132 align:middle line:90%
that this is true.

00:25:39.132 --> 00:25:40.340 align:middle line:90%
It really does work this way.

00:25:40.340 --> 00:25:42.132 align:middle line:84%
I'm going to show you
one piece of evidence

00:25:42.132 --> 00:25:43.683 align:middle line:84%
that this is how
wh- movement works.

00:25:43.683 --> 00:25:45.100 align:middle line:84%
It's going to be
kind of involved,

00:25:45.100 --> 00:25:47.610 align:middle line:84%
so I'll try to go
through it slowly.

00:25:47.610 --> 00:25:48.610 align:middle line:90%
Did you have a question?

00:25:48.610 --> 00:25:49.235 align:middle line:90%
AUDIENCE: Yeah.

00:25:49.235 --> 00:25:52.350 align:middle line:84%
So if we moved it up
to where the CP is,

00:25:52.350 --> 00:25:57.550 align:middle line:84%
then it would be like, "She will
say what that we should buy."

00:25:57.550 --> 00:25:59.830 align:middle line:90%
I feel like that sounds weird.

00:25:59.830 --> 00:26:04.197 align:middle line:84%
I would say like, "She will
say what we should buy,"

00:26:04.197 --> 00:26:04.780 align:middle line:90%
or something--

00:26:04.780 --> 00:26:05.270 align:middle line:90%
I don't know.

00:26:05.270 --> 00:26:05.860 align:middle line:90%
NORVIN RICHARDS: Yes.

00:26:05.860 --> 00:26:06.730 align:middle line:90%
No, you're right.

00:26:06.730 --> 00:26:11.123 align:middle line:84%
So "say," and we talked about--
there are verbs, like "wonder,"

00:26:11.123 --> 00:26:12.040 align:middle line:90%
that need to question.

00:26:12.040 --> 00:26:13.420 align:middle line:84%
There are verbs
like "think" that

00:26:13.420 --> 00:26:14.620 align:middle line:90%
need to not have a question.

00:26:14.620 --> 00:26:16.360 align:middle line:90%
"Say" can have either one.

00:26:16.360 --> 00:26:21.910 align:middle line:84%
So you can say, "She will
say that it is raining."

00:26:21.910 --> 00:26:26.470 align:middle line:90%
Or She will say who left."

00:26:26.470 --> 00:26:29.515 align:middle line:90%
So "say" has both options.

00:26:29.515 --> 00:26:31.390 align:middle line:84%
It can have either a
question or a statement.

00:26:31.390 --> 00:26:33.670 align:middle line:84%
But actually, the
evidence that I'm

00:26:33.670 --> 00:26:35.950 align:middle line:84%
going to show you about
successive cyclic movement

00:26:35.950 --> 00:26:39.250 align:middle line:84%
shows that it actually
doesn't matter

00:26:39.250 --> 00:26:45.250 align:middle line:84%
whether the higher verb can
select for a question or not.

00:26:45.250 --> 00:26:47.290 align:middle line:84%
That even for
verbs like "think,"

00:26:47.290 --> 00:26:49.210 align:middle line:84%
that have to select
for statements,

00:26:49.210 --> 00:26:51.650 align:middle line:84%
you get successive
cyclic movement.

00:26:51.650 --> 00:26:53.800 align:middle line:84%
So you're raising a very
good point, which is--

00:26:53.800 --> 00:26:56.500 align:middle line:84%
I'm sorry, I'm going
to hijack your point.

00:26:56.500 --> 00:26:58.660 align:middle line:84%
This movement, this
first movement,

00:26:58.660 --> 00:27:02.560 align:middle line:84%
it's not driven by the
need to form a question.

00:27:02.560 --> 00:27:04.630 align:middle line:84%
It's not that this wh-
phrase is moving here

00:27:04.630 --> 00:27:07.060 align:middle line:84%
to make that embedded
clause a question.

00:27:07.060 --> 00:27:09.088 align:middle line:90%
It's just moving there.

00:27:09.088 --> 00:27:11.380 align:middle line:84%
And then there's this question
about why, like why does

00:27:11.380 --> 00:27:13.270 align:middle line:84%
successive cyclic
movement happen.

00:27:13.270 --> 00:27:15.650 align:middle line:84%
But what I'll try to show you
now is that it does happen.

00:27:15.650 --> 00:27:18.150 align:middle line:84%
And so, we need to get used to
it and try to understand why.

00:27:18.150 --> 00:27:21.160 align:middle line:84%
But we need a theory of
why, because it's true.

00:27:21.160 --> 00:27:23.320 align:middle line:90%
It does happen.

00:27:23.320 --> 00:27:25.820 align:middle line:84%
I haven't yet shown you have
any evidence that this is true.

00:27:25.820 --> 00:27:28.420 align:middle line:84%
So far, all I'm
doing is asserting

00:27:28.420 --> 00:27:30.190 align:middle line:84%
through the force
of my personality

00:27:30.190 --> 00:27:32.440 align:middle line:84%
that this is how the
wh- movement works.

00:27:32.440 --> 00:27:34.210 align:middle line:84%
Let me now show
you some evidence.

00:27:34.210 --> 00:27:37.720 align:middle line:90%


00:27:37.720 --> 00:27:41.368 align:middle line:84%
The evidence involves--
do we have evidence

00:27:41.368 --> 00:27:42.160 align:middle line:90%
that this is right?

00:27:42.160 --> 00:27:43.110 align:middle line:90%
Yes.

00:27:43.110 --> 00:27:45.740 align:middle line:84%
The evidence comes-- there's
lots of evidence for this.

00:27:45.740 --> 00:27:47.440 align:middle line:84%
But I want to show you one of
my favorite pieces of evidence.

00:27:47.440 --> 00:27:49.030 align:middle line:84%
It comes from a
language called Dinka,

00:27:49.030 --> 00:27:50.350 align:middle line:90%
which is a Nilotic language.

00:27:50.350 --> 00:27:53.230 align:middle line:84%
It's spoken in South
Sudan and in lots

00:27:53.230 --> 00:27:55.690 align:middle line:84%
of diaspora communities,
including a big community

00:27:55.690 --> 00:28:00.120 align:middle line:84%
here in Boston, by
about 2 million people.

00:28:00.120 --> 00:28:02.200 align:middle line:90%
Its phonology is lots of fun.

00:28:02.200 --> 00:28:07.380 align:middle line:84%
We did a field methods class
on Dinka few years ago.

00:28:07.380 --> 00:28:11.700 align:middle line:84%
It has all kinds of shenanigans
it gets up to with vowels.

00:28:11.700 --> 00:28:16.560 align:middle line:84%
Actually, let me see, I can play
you these shenanigans in a way

00:28:16.560 --> 00:28:17.640 align:middle line:90%
that you can hear.

00:28:17.640 --> 00:28:19.410 align:middle line:84%
And I will see whether
this works or not.

00:28:19.410 --> 00:28:23.260 align:middle line:90%


00:28:23.260 --> 00:28:28.380 align:middle line:84%
So it has long and short
vowels, which itself, is not

00:28:28.380 --> 00:28:29.370 align:middle line:90%
all that unusual.

00:28:29.370 --> 00:28:32.340 align:middle line:84%
But it has three different
levels of vowel length.

00:28:32.340 --> 00:28:34.110 align:middle line:84%
So there are short
vowels and long vowels

00:28:34.110 --> 00:28:35.340 align:middle line:90%
and really long vowels.

00:28:35.340 --> 00:28:39.150 align:middle line:84%
So here are three
minimal triple.

00:28:39.150 --> 00:28:41.910 align:middle line:84%
You've got "mouse" and
"charcoal" and "pieces

00:28:41.910 --> 00:28:43.948 align:middle line:84%
of charcoal," both
singular and plural.

00:28:43.948 --> 00:28:45.240 align:middle line:90%
So here's the word for "mouse."

00:28:45.240 --> 00:28:48.430 align:middle line:84%
Let's see if I can play
it so you can hear it.

00:28:48.430 --> 00:28:48.990 align:middle line:90%
No, I can't.

00:28:48.990 --> 00:28:51.920 align:middle line:90%


00:28:51.920 --> 00:28:54.672 align:middle line:90%
[SPEECH SOUND]

00:28:54.672 --> 00:28:57.510 align:middle line:90%


00:28:57.510 --> 00:28:58.980 align:middle line:84%
Oh, I think I know
what happened.

00:28:58.980 --> 00:29:02.390 align:middle line:90%


00:29:02.390 --> 00:29:05.045 align:middle line:90%
I will put these--

00:29:05.045 --> 00:29:07.170 align:middle line:84%
[SPEECH SOUND] if
people will be quiet,

00:29:07.170 --> 00:29:08.420 align:middle line:90%
they may be able to hear this.

00:29:08.420 --> 00:29:11.390 align:middle line:84%
[SPEECH SOUND] And here
is the longer version.

00:29:11.390 --> 00:29:13.743 align:middle line:84%
[SPEECH SOUND] That's
a longer vowel.

00:29:13.743 --> 00:29:15.410 align:middle line:84%
And then we also,
somewhere around here,

00:29:15.410 --> 00:29:16.902 align:middle line:90%
have a really long vowel.

00:29:16.902 --> 00:29:18.682 align:middle line:90%
[SPEECH SOUND] There.

00:29:18.682 --> 00:29:20.390 align:middle line:84%
I'll put those sound
files on the website

00:29:20.390 --> 00:29:22.040 align:middle line:90%
so that you can hear them.

00:29:22.040 --> 00:29:24.208 align:middle line:84%
Basically, the word for
"mouse" is [SPEECH SOUND]..

00:29:24.208 --> 00:29:26.000 align:middle line:84%
The word for "charcoal"
is [SPEECH SOUND]..

00:29:26.000 --> 00:29:29.780 align:middle line:84%
And the word for pieces of
"charcoal" is [SPEECH SOUND]..

00:29:29.780 --> 00:29:33.290 align:middle line:90%
It also has tone.

00:29:33.290 --> 00:29:35.240 align:middle line:84%
And it has a contrast
between what are called

00:29:35.240 --> 00:29:37.580 align:middle line:90%
creaky and breathy voice.

00:29:37.580 --> 00:29:39.890 align:middle line:84%
And I'll play those sound
clips for you as well.

00:29:39.890 --> 00:29:44.450 align:middle line:84%
Actually, here, let's see
if I can get this to work.

00:29:44.450 --> 00:29:44.950 align:middle line:90%
No.

00:29:44.950 --> 00:29:45.450 align:middle line:90%
No.

00:29:45.450 --> 00:29:46.180 align:middle line:90%
No.

00:29:46.180 --> 00:29:47.170 align:middle line:90%
No.

00:29:47.170 --> 00:29:47.795 align:middle line:90%
No.

00:29:47.795 --> 00:29:48.295 align:middle line:90%
No.

00:29:48.295 --> 00:29:50.920 align:middle line:90%


00:29:50.920 --> 00:29:51.420 align:middle line:90%
I'm sorry.

00:29:51.420 --> 00:29:57.140 align:middle line:90%


00:29:57.140 --> 00:30:00.987 align:middle line:84%
My theory is that
this was not working

00:30:00.987 --> 00:30:03.570 align:middle line:84%
because I didn't have it plugged
in when I opened the program.

00:30:03.570 --> 00:30:04.820 align:middle line:90%
Let's see whether that's true.

00:30:04.820 --> 00:30:08.050 align:middle line:90%


00:30:08.050 --> 00:30:08.550 align:middle line:90%
Doo.

00:30:08.550 --> 00:30:08.640 align:middle line:90%
Doo.

00:30:08.640 --> 00:30:08.790 align:middle line:90%
Doo.

00:30:08.790 --> 00:30:08.940 align:middle line:90%
Doo.

00:30:08.940 --> 00:30:09.440 align:middle line:90%
Doo.

00:30:09.440 --> 00:30:11.990 align:middle line:90%


00:30:11.990 --> 00:30:13.678 align:middle line:90%
Ah.

00:30:13.678 --> 00:30:15.710 align:middle line:90%
[SPEECH SOUND] There we go.

00:30:15.710 --> 00:30:17.630 align:middle line:84%
[SPEECH SOUND] So
that's creaky voice.

00:30:17.630 --> 00:30:19.388 align:middle line:90%
That's the word for "berries."

00:30:19.388 --> 00:30:20.555 align:middle line:90%
Here's the word for "women."

00:30:20.555 --> 00:30:25.850 align:middle line:90%


00:30:25.850 --> 00:30:28.460 align:middle line:90%
[SPEECH SOUND] [SPEECH SOUND].

00:30:28.460 --> 00:30:31.192 align:middle line:84%
So you do fancy things
with your vocal cords

00:30:31.192 --> 00:30:32.400 align:middle line:90%
when you speak this language.

00:30:32.400 --> 00:30:35.525 align:middle line:84%
You either have to have-- it
has different dialects of Dinka

00:30:35.525 --> 00:30:39.350 align:middle line:84%
have different numbers
of voice qualities.

00:30:39.350 --> 00:30:41.120 align:middle line:84%
But in this dialect,
there are two.

00:30:41.120 --> 00:30:44.150 align:middle line:84%
There's creaky voice
where your vocal cords

00:30:44.150 --> 00:30:47.450 align:middle line:84%
are under a lot of tension,
I guess, as you're speaking.

00:30:47.450 --> 00:30:48.745 align:middle line:90%
[SPEECH SOUND] [VOWEL SOUND].

00:30:48.745 --> 00:30:50.120 align:middle line:84%
So that's the word
for "berries."

00:30:50.120 --> 00:30:51.800 align:middle line:90%
Here's the word for "women."

00:30:51.800 --> 00:30:55.850 align:middle line:84%
[SPEECH SOUND] [SPEECH SOUND]
where the voice is breathy.

00:30:55.850 --> 00:30:59.300 align:middle line:84%
There's actually a more
minimal pair in here somewhere,

00:30:59.300 --> 00:31:00.760 align:middle line:90%
but I'm not finding it.

00:31:00.760 --> 00:31:02.090 align:middle line:90%
Never mind.

00:31:02.090 --> 00:31:04.610 align:middle line:84%
So that's what
Dinka sounds like.

00:31:04.610 --> 00:31:08.240 align:middle line:84%
It is lots of fun
to try to study,

00:31:08.240 --> 00:31:13.520 align:middle line:84%
partly because a lot of
its morphology is tonal,

00:31:13.520 --> 00:31:17.420 align:middle line:84%
involves changes to tone or
to creaky or breathy voice,

00:31:17.420 --> 00:31:18.500 align:middle line:90%
or to vowel length.

00:31:18.500 --> 00:31:21.080 align:middle line:84%
So you can see, for example,
that the word for charcoal

00:31:21.080 --> 00:31:23.570 align:middle line:84%
forms its plural by
lengthening the vowel, which

00:31:23.570 --> 00:31:24.440 align:middle line:90%
is already long.

00:31:24.440 --> 00:31:26.060 align:middle line:84%
And also making
it low tone, so it

00:31:26.060 --> 00:31:30.770 align:middle line:84%
changes from [SPEECH SOUND]
to [SPEECH SOUND]..

00:31:30.770 --> 00:31:33.305 align:middle line:84%
We lived in fear that we were
missing half of the morphemes.

00:31:33.305 --> 00:31:35.850 align:middle line:90%
[LAUGHING]

00:31:35.850 --> 00:31:37.490 align:middle line:84%
So that's enough
about Dinka phonology.

00:31:37.490 --> 00:31:40.450 align:middle line:84%
Let me tell you something
things about Dinka syntax.

00:31:40.450 --> 00:31:43.490 align:middle line:90%
Dinka is V2 language.

00:31:43.490 --> 00:31:46.330 align:middle line:84%
So you can see
some examples here.

00:31:46.330 --> 00:31:48.640 align:middle line:84%
If you want to say
things like "Can

00:31:48.640 --> 00:31:52.822 align:middle line:84%
will buy Bol some
clothes in the town,"

00:31:52.822 --> 00:31:54.280 align:middle line:84%
there's a position
at the beginning

00:31:54.280 --> 00:31:57.280 align:middle line:84%
of the sentence that has to
get filled by some phrase.

00:31:57.280 --> 00:32:00.583 align:middle line:84%
It can be filled by "Can" or
by "town," or by "clothes."

00:32:00.583 --> 00:32:02.500 align:middle line:84%
You can pick any of these
things and move them

00:32:02.500 --> 00:32:04.600 align:middle line:90%
into the first position.

00:32:04.600 --> 00:32:07.060 align:middle line:84%
That phrase is then
followed by an auxiliary

00:32:07.060 --> 00:32:08.140 align:middle line:90%
in all of these examples.

00:32:08.140 --> 00:32:09.682 align:middle line:84%
If there isn't an
auxiliary, then you

00:32:09.682 --> 00:32:10.998 align:middle line:90%
get the verb in that position.

00:32:10.998 --> 00:32:12.790 align:middle line:84%
And then you have the
rest of the sentence.

00:32:12.790 --> 00:32:16.900 align:middle line:84%
Dinka is unlike German
in one or two ways.

00:32:16.900 --> 00:32:19.780 align:middle line:84%
German doesn't have creaky and
breathy voice, for one thing.

00:32:19.780 --> 00:32:22.780 align:middle line:84%
But also, the verb
in German, we saw,

00:32:22.780 --> 00:32:24.850 align:middle line:84%
goes at the end of
the verb phrase.

00:32:24.850 --> 00:32:27.610 align:middle line:84%
The verb in Dinka goes
earlier in the verb phrases.

00:32:27.610 --> 00:32:28.660 align:middle line:90%
It's head initial.

00:32:28.660 --> 00:32:31.660 align:middle line:84%
It's followed by its complement,
with some complications,

00:32:31.660 --> 00:32:32.720 align:middle line:90%
as we'll see.

00:32:32.720 --> 00:32:38.590 align:middle line:84%
So you can say "Can will Bol
buy clothes in the town,"

00:32:38.590 --> 00:32:42.250 align:middle line:84%
or you can say "Clothes will
Cán Bol buy in the town,"

00:32:42.250 --> 00:32:43.420 align:middle line:90%
or you can say "town"--

00:32:43.420 --> 00:32:44.795 align:middle line:84%
you can take any
of these phrases

00:32:44.795 --> 00:32:46.840 align:middle line:84%
and move them into
first position.

00:32:46.840 --> 00:32:48.340 align:middle line:84%
Dinka is also
different from German

00:32:48.340 --> 00:32:50.320 align:middle line:84%
in that there's a
morpheme in Dinka that

00:32:50.320 --> 00:32:52.780 align:middle line:84%
tells you the number
of the thing you've

00:32:52.780 --> 00:32:54.080 align:middle line:90%
put in first position.

00:32:54.080 --> 00:32:57.940 align:middle line:84%
So you can see
that the auxiliary

00:32:57.940 --> 00:33:03.400 align:middle line:84%
in the first and the third
examples starts with ah.

00:33:03.400 --> 00:33:07.240 align:middle line:84%
And in the second
example, starts with ah.

00:33:07.240 --> 00:33:09.130 align:middle line:84%
So the ah, in the
second example,

00:33:09.130 --> 00:33:10.975 align:middle line:84%
is indicating that the
clothes are plural,

00:33:10.975 --> 00:33:13.780 align:middle line:84%
that the thing in first
position is plural.

00:33:13.780 --> 00:33:15.743 align:middle line:90%
So German doesn't have that.

00:33:15.743 --> 00:33:18.160 align:middle line:84%
There are other alternations
that you can see there, which

00:33:18.160 --> 00:33:19.680 align:middle line:90%
I won't try to talk about.

00:33:19.680 --> 00:33:20.333 align:middle line:90%
Oh, well.

00:33:20.333 --> 00:33:22.000 align:middle line:84%
So you can also see
there's a difference

00:33:22.000 --> 00:33:24.950 align:middle line:84%
between the auxiliary
and the first example,

00:33:24.950 --> 00:33:26.290 align:middle line:90%
which is [SPEECH SOUND].

00:33:26.290 --> 00:33:28.720 align:middle line:84%
That tells you that the
thing in first position

00:33:28.720 --> 00:33:31.252 align:middle line:84%
is the subject,
whereas the auxiliary

00:33:31.252 --> 00:33:32.710 align:middle line:84%
in the second and
third examples is

00:33:32.710 --> 00:33:34.840 align:middle line:84%
B is telling you that the
thing in first position

00:33:34.840 --> 00:33:35.870 align:middle line:90%
is not the subject.

00:33:35.870 --> 00:33:37.995 align:middle line:84%
So there's all this
morphology, and as you can see,

00:33:37.995 --> 00:33:40.450 align:middle line:90%
it's tricky to keep track of.

00:33:40.450 --> 00:33:43.310 align:middle line:90%


00:33:43.310 --> 00:33:45.130 align:middle line:90%
So it's V2.

00:33:45.130 --> 00:33:47.410 align:middle line:84%
The last example
demonstrates this.

00:33:47.410 --> 00:33:51.310 align:middle line:84%
So it's ungrammatical to just
have nothing in first position.

00:33:51.310 --> 00:33:53.800 align:middle line:84%
You have to put something
in first position.

00:33:53.800 --> 00:33:54.415 align:middle line:90%
V2 language.

00:33:54.415 --> 00:33:56.960 align:middle line:90%


00:33:56.960 --> 00:33:59.660 align:middle line:84%
Now there is another
position in the Dinka clause

00:33:59.660 --> 00:34:00.920 align:middle line:90%
that has to get occupied.

00:34:00.920 --> 00:34:06.260 align:middle line:84%
So if you want to say in Dinka,
"I saw a giraffe," you can say,

00:34:06.260 --> 00:34:08.330 align:middle line:90%
[SPEAKING DINKA].

00:34:08.330 --> 00:34:14.120 align:middle line:84%
No, sorry, [SPEAKING DINKA],,
so C is breathy.

00:34:14.120 --> 00:34:17.420 align:middle line:84%
You've got "I," the word for
"I," in first position here,

00:34:17.420 --> 00:34:19.159 align:middle line:84%
it is satisfying
the V2 requirement.

00:34:19.159 --> 00:34:21.139 align:middle line:84%
And throughout,
I'm going to make

00:34:21.139 --> 00:34:23.400 align:middle line:84%
things in the V2 requirement,
things in the specifier

00:34:23.400 --> 00:34:25.572 align:middle line:84%
or CP, things satisfying
that need for there

00:34:25.572 --> 00:34:26.989 align:middle line:84%
to be something
in first position,

00:34:26.989 --> 00:34:28.820 align:middle line:90%
I'll make those things blue.

00:34:28.820 --> 00:34:32.960 align:middle line:84%
There is another position that
has to get occupied though.

00:34:32.960 --> 00:34:36.170 align:middle line:84%
So you can say "I
(auxiliary) giraffe saw."

00:34:36.170 --> 00:34:38.120 align:middle line:84%
That's how you say
"I saw a giraffe."

00:34:38.120 --> 00:34:40.504 align:middle line:84%
You can't leave
"giraffe" after the verb.

00:34:40.504 --> 00:34:43.080 align:middle line:90%


00:34:43.080 --> 00:34:45.650 align:middle line:84%
If you-- basically,
the generalization

00:34:45.650 --> 00:34:48.800 align:middle line:84%
is that if there is
a deep, an NP object,

00:34:48.800 --> 00:34:51.050 align:middle line:84%
the NP object has to
go before the verb.

00:34:51.050 --> 00:34:52.670 align:middle line:84%
Everything else
goes after the verb.

00:34:52.670 --> 00:34:55.440 align:middle line:84%
But the NP object
goes before the verb.

00:34:55.440 --> 00:34:57.200 align:middle line:84%
So there's another
position down there.

00:34:57.200 --> 00:34:59.537 align:middle line:84%
I won't try to talk
about how we figure out

00:34:59.537 --> 00:35:00.620 align:middle line:90%
where that is in the tree.

00:35:00.620 --> 00:35:02.662 align:middle line:84%
That won't matter for what
I'm about to show you.

00:35:02.662 --> 00:35:04.903 align:middle line:84%
But I want you to
know about it, partly

00:35:04.903 --> 00:35:07.070 align:middle line:84%
because it will make things
slightly less confusing,

00:35:07.070 --> 00:35:08.300 align:middle line:90%
I hope.

00:35:08.300 --> 00:35:10.610 align:middle line:84%
If you have two objects,
like you say things like,

00:35:10.610 --> 00:35:15.260 align:middle line:84%
"I gave Ayen a book," you
can choose either of them

00:35:15.260 --> 00:35:16.560 align:middle line:90%
to go in first position.

00:35:16.560 --> 00:35:20.960 align:middle line:84%
So you can say either, "I
have a book given Ayen,"

00:35:20.960 --> 00:35:23.210 align:middle line:90%
or "I have Ayen given the book."

00:35:23.210 --> 00:35:25.700 align:middle line:84%
One of those two objects
has to go before "give."

00:35:25.700 --> 00:35:27.440 align:middle line:84%
You can't leave them
both after "give."

00:35:27.440 --> 00:35:29.440 align:middle line:84%
I don't think I have this
anywhere on the slide,

00:35:29.440 --> 00:35:32.360 align:middle line:84%
but you also can't put
them both before "give."

00:35:32.360 --> 00:35:33.297 align:middle line:90%
So it's a V2 language.

00:35:33.297 --> 00:35:34.880 align:middle line:84%
Something has to be
in first position.

00:35:34.880 --> 00:35:37.190 align:middle line:84%
And there's another position
right before the verb

00:35:37.190 --> 00:35:39.080 align:middle line:84%
that also has to get
occupied, if there

00:35:39.080 --> 00:35:40.393 align:middle line:90%
is a noun-phrase object.

00:35:40.393 --> 00:35:41.810 align:middle line:84%
If there is no
noun-phrase object,

00:35:41.810 --> 00:35:45.470 align:middle line:84%
then nothing goes in
that second position.

00:35:45.470 --> 00:35:48.506 align:middle line:84%
That lower position,
the red position.

00:35:48.506 --> 00:35:52.040 align:middle line:90%
OK so far?

00:35:52.040 --> 00:35:54.000 align:middle line:90%
This is all set up.

00:35:54.000 --> 00:35:57.420 align:middle line:84%
So Dinka has two positions
in the clause that

00:35:57.420 --> 00:35:59.160 align:middle line:90%
absolutely have to be filled.

00:35:59.160 --> 00:36:02.790 align:middle line:84%
There's the specifier of
CP, so it's a V2 language.

00:36:02.790 --> 00:36:05.970 align:middle line:84%
The blue position, the
position at the beginning

00:36:05.970 --> 00:36:08.520 align:middle line:84%
of every clause, has to
get filled by some phrase.

00:36:08.520 --> 00:36:10.470 align:middle line:84%
And Dinka is V2,
so it can be filled

00:36:10.470 --> 00:36:12.450 align:middle line:84%
by a variety of
different phrases.

00:36:12.450 --> 00:36:13.950 align:middle line:84%
And then there is
another position

00:36:13.950 --> 00:36:15.992 align:middle line:84%
that's right before the
verb that also absolutely

00:36:15.992 --> 00:36:17.410 align:middle line:90%
has to get filled.

00:36:17.410 --> 00:36:20.364 align:middle line:84%
So if you say in
Dinka, "Yaar told Ayen

00:36:20.364 --> 00:36:23.760 align:middle line:84%
that Bol sent Deng
to the cattle camp"--

00:36:23.760 --> 00:36:26.490 align:middle line:84%
The cattle camp is
where the cattle are.

00:36:26.490 --> 00:36:27.990 align:middle line:84%
Young people go to
the cattle camp.

00:36:27.990 --> 00:36:29.790 align:middle line:84%
Our consultant told us
that the cattle camp

00:36:29.790 --> 00:36:31.665 align:middle line:84%
is where young people
get to know each other.

00:36:31.665 --> 00:36:35.100 align:middle line:84%
So men and women mostly don't
mix, but at the cattle camp,

00:36:35.100 --> 00:36:35.760 align:middle line:90%
then they do.

00:36:35.760 --> 00:36:38.370 align:middle line:90%


00:36:38.370 --> 00:36:41.720 align:middle line:90%
So Yaar and Ayen--

00:36:41.720 --> 00:36:44.670 align:middle line:84%
I'm trying to remember, "Yaar"
and "Ayen" are women's names.

00:36:44.670 --> 00:36:46.710 align:middle line:84%
And "Bol" and "Deng"
are men's names.

00:36:46.710 --> 00:36:49.380 align:middle line:84%
"Bol"-- I think "Bol"
means "a person who

00:36:49.380 --> 00:36:51.570 align:middle line:90%
is born after a set of twins."

00:36:51.570 --> 00:36:53.220 align:middle line:84%
It's apparently a
really common name.

00:36:53.220 --> 00:36:56.970 align:middle line:84%
Twins, I guess, are really
common among the Dinka.

00:36:56.970 --> 00:37:00.960 align:middle line:84%
"Ayen" is a name for
a color, just like a--

00:37:00.960 --> 00:37:03.270 align:middle line:90%
yeah, anyway.

00:37:03.270 --> 00:37:05.700 align:middle line:90%
So here's a sentence.

00:37:05.700 --> 00:37:08.190 align:middle line:84%
The main clause and
the embedded clause

00:37:08.190 --> 00:37:10.570 align:middle line:84%
both have their subjects
in the first position.

00:37:10.570 --> 00:37:13.620 align:middle line:84%
So you've got Yaar before the
auxiliary in the main clause

00:37:13.620 --> 00:37:15.660 align:middle line:84%
and you've got Bol
before the auxiliary

00:37:15.660 --> 00:37:18.910 align:middle line:90%
in the embedded clause.

00:37:18.910 --> 00:37:22.060 align:middle line:84%
And in the main
clause, you've got

00:37:22.060 --> 00:37:23.950 align:middle line:84%
Ayen, the person
who's being told,

00:37:23.950 --> 00:37:27.843 align:middle line:84%
going before the verb "told,"
following that red position.

00:37:27.843 --> 00:37:29.260 align:middle line:84%
And in the embedded
clause, you've

00:37:29.260 --> 00:37:34.960 align:middle line:84%
got Deng, the person who's
being sent, preceding the verb,

00:37:34.960 --> 00:37:36.685 align:middle line:90%
filling that red position.

00:37:36.685 --> 00:37:39.010 align:middle line:90%
OK so far?

00:37:39.010 --> 00:37:41.860 align:middle line:90%
So now a little bit of Dinka.

00:37:41.860 --> 00:37:46.060 align:middle line:84%
Here's the thing,
if you wh- extract--

00:37:46.060 --> 00:37:50.935 align:middle line:84%
so suppose you want to ask,
"Who did Yaar tell Ayen that Bol

00:37:50.935 --> 00:37:52.060 align:middle line:90%
sent to the cattle camp?"--

00:37:52.060 --> 00:37:55.060 align:middle line:84%
So here we've changed
Deng into a wh- word,

00:37:55.060 --> 00:37:59.200 align:middle line:84%
and we're going to move
Deng into the main clause.

00:37:59.200 --> 00:38:01.840 align:middle line:84%
So the object of "send"
is going to be a wh- word

00:38:01.840 --> 00:38:03.760 align:middle line:84%
and it's going to go
in the main clause.

00:38:03.760 --> 00:38:06.020 align:middle line:84%
The whole thing is
going to be a question.

00:38:06.020 --> 00:38:07.040 align:middle line:90%
Here's the observation.

00:38:07.040 --> 00:38:09.650 align:middle line:84%
If you do that, all
of those positions

00:38:09.650 --> 00:38:14.240 align:middle line:84%
that normally absolutely have
to be filled have to be empty.

00:38:14.240 --> 00:38:20.220 align:middle line:84%
So "who" started off as
the object of "send."

00:38:20.220 --> 00:38:22.910 align:middle line:84%
And so that rightmost
red position,

00:38:22.910 --> 00:38:25.048 align:middle line:84%
the one that "Deng"
used to be in, is empty.

00:38:25.048 --> 00:38:26.090 align:middle line:90%
That's not so surprising.

00:38:26.090 --> 00:38:29.360 align:middle line:90%
That's where "who" came from.

00:38:29.360 --> 00:38:31.850 align:middle line:84%
But what's interesting is
that the blue position right

00:38:31.850 --> 00:38:33.680 align:middle line:84%
above that, the one
before the auxiliary

00:38:33.680 --> 00:38:36.410 align:middle line:84%
of the embedded clause,
the one that's normally--

00:38:36.410 --> 00:38:39.470 align:middle line:84%
that was filled by "Bol"
in the first example,

00:38:39.470 --> 00:38:41.840 align:middle line:90%
can't be filled by "Bol."

00:38:41.840 --> 00:38:44.510 align:middle line:84%
"Bol" has to not
be there anymore.

00:38:44.510 --> 00:38:48.140 align:middle line:84%
And similarly, the red
position before "tell,"

00:38:48.140 --> 00:38:50.900 align:middle line:84%
the first red position in
the second example there,

00:38:50.900 --> 00:38:53.750 align:middle line:84%
the one that's an empty
box, has to be empty.

00:38:53.750 --> 00:38:55.857 align:middle line:84%
Normally it would
be filled by "Ayen."

00:38:55.857 --> 00:38:58.190 align:middle line:84%
That's what's it's being
filled by in the first example.

00:38:58.190 --> 00:39:02.880 align:middle line:84%
But if you wh- move across
it, it has to be empty.

00:39:02.880 --> 00:39:05.150 align:middle line:84%
So these positions that
ordinarily absolutely,

00:39:05.150 --> 00:39:08.000 align:middle line:84%
absolutely have to be filled,
cannot leave them empty,

00:39:08.000 --> 00:39:10.430 align:middle line:84%
suddenly they have to be
empty if you're moving a wh-

00:39:10.430 --> 00:39:12.650 align:middle line:90%
phrase across them.

00:39:12.650 --> 00:39:17.710 align:middle line:84%
Well we can understand
why if we're

00:39:17.710 --> 00:39:20.230 align:middle line:84%
willing to believe in
successive cyclic movement.

00:39:20.230 --> 00:39:24.280 align:middle line:84%
So we just have to be willing
to believe the wh- movement is

00:39:24.280 --> 00:39:28.210 align:middle line:84%
obliged to move not
in one mighty leap,

00:39:28.210 --> 00:39:30.970 align:middle line:90%
but in a series of little hops.

00:39:30.970 --> 00:39:33.280 align:middle line:84%
It has to land in all
of these open positions.

00:39:33.280 --> 00:39:35.140 align:middle line:84%
And that's why they
can't be filled the way

00:39:35.140 --> 00:39:38.410 align:middle line:84%
they normally would be
if you're doing long wh-

00:39:38.410 --> 00:39:39.350 align:middle line:90%
movement across them.

00:39:39.350 --> 00:39:41.680 align:middle line:84%
So there's an explanation
for why these positions are

00:39:41.680 --> 00:39:43.513 align:middle line:90%
emptied out.

00:39:43.513 --> 00:39:45.430 align:middle line:84%
I'll show you one other
cool thing about Dinka

00:39:45.430 --> 00:39:48.880 align:middle line:84%
and then I think
we're done with Dinka.

00:39:48.880 --> 00:39:52.570 align:middle line:84%
Dinka has a distinction
between singular and plural

00:39:52.570 --> 00:39:55.030 align:middle line:90%
versions of wh- words.

00:39:55.030 --> 00:39:57.940 align:middle line:84%
So there's "yeŋà" which
is the word for "who,"

00:39:57.940 --> 00:40:01.300 align:middle line:84%
and then there's "yèyíŋà," which
is the word for "who" plural,

00:40:01.300 --> 00:40:01.855 align:middle line:90%
"who all."

00:40:01.855 --> 00:40:03.730 align:middle line:84%
English doesn't have a
word for "who" plural,

00:40:03.730 --> 00:40:07.060 align:middle line:84%
but plenty of languages
do, and Dinka is one.

00:40:07.060 --> 00:40:12.340 align:middle line:84%
If you ask a plural
question, you

00:40:12.340 --> 00:40:15.340 align:middle line:84%
put that "who" at the
beginning in the spec of CP,

00:40:15.340 --> 00:40:17.480 align:middle line:90%
of the clause that you're in.

00:40:17.480 --> 00:40:20.020 align:middle line:84%
But you're required to
leave behind the plural

00:40:20.020 --> 00:40:24.860 align:middle line:84%
morpheme in the red positions,
the positions before the verb.

00:40:24.860 --> 00:40:26.822 align:middle line:90%
So "Who all did Bol see?"

00:40:26.822 --> 00:40:28.780 align:middle line:84%
You've got this "who"
plural in first position.

00:40:28.780 --> 00:40:31.820 align:middle line:84%
But you've also got this "ké"
that goes in the red position,

00:40:31.820 --> 00:40:33.320 align:middle line:84%
the position that's
before the verb.

00:40:33.320 --> 00:40:36.220 align:middle line:84%
If you wh- extract something
from the verb phrase,

00:40:36.220 --> 00:40:39.220 align:middle line:84%
like an object, you have to
leave this plural morpheme

00:40:39.220 --> 00:40:41.020 align:middle line:90%
behind.

00:40:41.020 --> 00:40:44.830 align:middle line:84%
So interesting fact, if
you do a longer question

00:40:44.830 --> 00:40:47.830 align:middle line:84%
than that, like "Who all do
you think that Bol saw?"--

00:40:47.830 --> 00:40:49.720 align:middle line:84%
So you've got "who"
plural at the beginning,

00:40:49.720 --> 00:40:52.900 align:middle line:84%
and you've got "ké," in both
of the red positions along

00:40:52.900 --> 00:40:54.340 align:middle line:90%
the way.

00:40:54.340 --> 00:40:56.380 align:middle line:84%
As you wh- move up
the clause, it's

00:40:56.380 --> 00:40:58.670 align:middle line:84%
like you leave
behind this little--

00:40:58.670 --> 00:41:03.850 align:middle line:84%
this was plural thing,
as you move along.

00:41:03.850 --> 00:41:07.840 align:middle line:84%
So you get multiple
"ké"s as you go up.

00:41:07.840 --> 00:41:13.537 align:middle line:84%
"Ké" sort of resembles the
third-person plural pronoun,

00:41:13.537 --> 00:41:14.620 align:middle line:90%
which is [SPEAKING DINKA].

00:41:14.620 --> 00:41:16.720 align:middle line:90%
So that's the word for they.

00:41:16.720 --> 00:41:19.460 align:middle line:84%
So we figure it's
somehow related to this.

00:41:19.460 --> 00:41:23.350 align:middle line:84%
So Dinka gives us some evidence
for this idea, this wacky idea

00:41:23.350 --> 00:41:27.060 align:middle line:84%
of Chomsky's from the
'70's that wh- movement has

00:41:27.060 --> 00:41:29.560 align:middle line:84%
to be successive cyclic, that
it's not possible to just move

00:41:29.560 --> 00:41:30.227 align:middle line:90%
in one big jump.

00:41:30.227 --> 00:41:32.500 align:middle line:84%
You have to move in a
bunch of little hops.

00:41:32.500 --> 00:41:36.250 align:middle line:84%
So Dinka has a
detector for landing

00:41:36.250 --> 00:41:38.500 align:middle line:84%
in positions-- two
detectors, actually.

00:41:38.500 --> 00:41:39.932 align:middle line:90%
One is this "ké" thing.

00:41:39.932 --> 00:41:41.890 align:middle line:84%
And the other is just
whether certain positions

00:41:41.890 --> 00:41:43.520 align:middle line:90%
are filled or empty.

00:41:43.520 --> 00:41:46.097 align:middle line:84%
So these positions that
normally have to be filled,

00:41:46.097 --> 00:41:47.680 align:middle line:84%
have to empty
themselves out if you're

00:41:47.680 --> 00:41:50.110 align:middle line:84%
going to wh- move through
them, by hypothesis,

00:41:50.110 --> 00:41:52.360 align:middle line:84%
because the wh- word
is landing in them

00:41:52.360 --> 00:41:55.480 align:middle line:90%
and moving on the way up.

00:41:55.480 --> 00:41:56.590 align:middle line:90%
One piece of evidence.

00:41:56.590 --> 00:41:58.750 align:middle line:84%
There are zillions of
pieces of evidence for this.

00:41:58.750 --> 00:42:00.140 align:middle line:90%
This is one of my favorites.

00:42:00.140 --> 00:42:04.090 align:middle line:90%


00:42:04.090 --> 00:42:06.600 align:middle line:90%
Now we are done with syntax.

00:42:06.600 --> 00:42:09.967 align:middle line:84%
Let us have a moment
of silence for syntax.

00:42:09.967 --> 00:42:11.800 align:middle line:84%
Does anybody have any
questions about syntax

00:42:11.800 --> 00:42:14.080 align:middle line:84%
before we leave syntax
behind and begin

00:42:14.080 --> 00:42:16.040 align:middle line:90%
talking about something else?

00:42:16.040 --> 00:42:20.290 align:middle line:84%
So if you are looking
at the syllabus,

00:42:20.290 --> 00:42:23.230 align:middle line:84%
there was going to
be a day on dialect.

00:42:23.230 --> 00:42:25.277 align:middle line:84%
I will move that day to
the end of semantics.

00:42:25.277 --> 00:42:27.610 align:middle line:84%
So we're now, more or less,
caught up with the syllabus.

00:42:27.610 --> 00:42:29.568 align:middle line:84%
We're going to spend some
time doing semantics.

00:42:29.568 --> 00:42:32.080 align:middle line:84%
And then, we will have
a grab bag of topics

00:42:32.080 --> 00:42:35.080 align:middle line:84%
after we're done with
semantics, on various things.

00:42:35.080 --> 00:42:37.480 align:middle line:84%
Any questions about syntax
before we leave syntax behind?

00:42:37.480 --> 00:42:40.320 align:middle line:90%


00:42:40.320 --> 00:42:41.910 align:middle line:90%
All right.

00:42:41.910 --> 00:42:46.030 align:middle line:84%
Let us start in
on semantics then.

00:42:46.030 --> 00:42:49.180 align:middle line:84%
So we've done morphology,
which is the building

00:42:49.180 --> 00:42:51.280 align:middle line:90%
up of words out of morphemes.

00:42:51.280 --> 00:42:53.770 align:middle line:84%
We've done syntax, which is
the building of a sentences out

00:42:53.770 --> 00:42:54.920 align:middle line:90%
of words.

00:42:54.920 --> 00:42:56.830 align:middle line:84%
We're now going to
do semantics, which

00:42:56.830 --> 00:43:00.220 align:middle line:84%
is the study of how
you make meanings out

00:43:00.220 --> 00:43:02.800 align:middle line:90%
of smaller meanings.

00:43:02.800 --> 00:43:06.350 align:middle line:84%
That's what semantics
is all about.

00:43:06.350 --> 00:43:08.710 align:middle line:84%
There's a going
hypothesis, which is that--

00:43:08.710 --> 00:43:11.830 align:middle line:84%
something like if you
completely understand

00:43:11.830 --> 00:43:14.470 align:middle line:84%
the meanings of the various
parts of the sentence,

00:43:14.470 --> 00:43:16.180 align:middle line:84%
the various morphemes
that combine,

00:43:16.180 --> 00:43:18.970 align:middle line:84%
that there ought to be a
simple set of rules for how

00:43:18.970 --> 00:43:21.250 align:middle line:84%
to combine meanings,
that would give you

00:43:21.250 --> 00:43:22.420 align:middle line:90%
the meaning of the sentence.

00:43:22.420 --> 00:43:24.545 align:middle line:84%
It's called-- this idea is
called compositionality.

00:43:24.545 --> 00:43:28.270 align:middle line:84%
It's the idea that we should
pay a lot of attention

00:43:28.270 --> 00:43:32.170 align:middle line:84%
to specifying very completely
what particular morphemes mean.

00:43:32.170 --> 00:43:35.680 align:middle line:84%
That if we'll do that, we can
have simple rules to generate

00:43:35.680 --> 00:43:37.420 align:middle line:90%
the meanings of sentences.

00:43:37.420 --> 00:43:39.370 align:middle line:84%
So we're going to
do some semantics.

00:43:39.370 --> 00:43:41.380 align:middle line:84%
I'm going introduce
you to semantics.

00:43:41.380 --> 00:43:43.720 align:middle line:84%
And semantics is going to be
like the other topics we've

00:43:43.720 --> 00:43:45.303 align:middle line:84%
talked about in this
class, in that, I

00:43:45.303 --> 00:43:48.010 align:middle line:84%
hope to tell you enough about
it to make it sound interesting.

00:43:48.010 --> 00:43:53.290 align:middle line:84%
And then I will abandon you and
let you study it on your own.

00:43:53.290 --> 00:43:56.140 align:middle line:84%
Go take-- there is an
undergrad intro syntax class--

00:43:56.140 --> 00:43:58.697 align:middle line:84%
semantics class that I
encourage you to take,

00:43:58.697 --> 00:44:00.280 align:middle line:84%
if you find the stuff
that we're going

00:44:00.280 --> 00:44:01.990 align:middle line:90%
to talk about interesting.

00:44:01.990 --> 00:44:05.080 align:middle line:84%
There's lots of interesting
work in semantics.

00:44:05.080 --> 00:44:09.520 align:middle line:84%
So let's talk about some meaning
relations, some classics,

00:44:09.520 --> 00:44:12.850 align:middle line:84%
like meaning relations
involving words.

00:44:12.850 --> 00:44:14.800 align:middle line:84%
So we can say that words
are synonyms, if they

00:44:14.800 --> 00:44:16.080 align:middle line:90%
seem to mean the same thing.

00:44:16.080 --> 00:44:18.400 align:middle line:84%
So words like
"purchase" and "buy."

00:44:18.400 --> 00:44:20.110 align:middle line:84%
And that words are
antonyms if they

00:44:20.110 --> 00:44:23.620 align:middle line:90%
seem to mean opposite things.

00:44:23.620 --> 00:44:25.090 align:middle line:90%
This is an antiquated example.

00:44:25.090 --> 00:44:26.020 align:middle line:90%
I should fix this.

00:44:26.020 --> 00:44:28.780 align:middle line:90%
So "male" and "female."

00:44:28.780 --> 00:44:31.370 align:middle line:84%
But there are fancier
things to say about meaning.

00:44:31.370 --> 00:44:34.180 align:middle line:84%
So let's consider the
various ways someone might

00:44:34.180 --> 00:44:35.830 align:middle line:90%
want to refer to me.

00:44:35.830 --> 00:44:39.228 align:middle line:84%
So people usually say at
the beginning of classes--

00:44:39.228 --> 00:44:41.020 align:middle line:84%
"It doesn't matter to
me what you call me."

00:44:41.020 --> 00:44:44.110 align:middle line:84%
Some people call me Professor
Richards or call me Professor.

00:44:44.110 --> 00:44:48.040 align:middle line:84%
Some people call me Norvin,
which is my first name.

00:44:48.040 --> 00:44:49.848 align:middle line:84%
If you can't remember
my name, there

00:44:49.848 --> 00:44:51.640 align:middle line:84%
are other things you
might want to call me.

00:44:51.640 --> 00:44:52.140 align:middle line:90%
[LAUGHING]

00:44:52.140 --> 00:44:52.725 align:middle line:90%
All right.

00:44:52.725 --> 00:44:55.840 align:middle line:84%
There's a variety of things
that you might say about me,

00:44:55.840 --> 00:44:59.290 align:middle line:90%
depending on the circumstances.

00:44:59.290 --> 00:45:03.100 align:middle line:84%
And to say that these
are all things that you

00:45:03.100 --> 00:45:07.840 align:middle line:84%
can use to refer to me is to
say that it's usually true

00:45:07.840 --> 00:45:10.570 align:middle line:84%
if you have a
sentence and it has

00:45:10.570 --> 00:45:14.590 align:middle line:84%
one of those expressions
in it, if you change just

00:45:14.590 --> 00:45:16.600 align:middle line:84%
that expression to
another expression,

00:45:16.600 --> 00:45:19.900 align:middle line:84%
that the sentence, the meaning
of the sentence won't change.

00:45:19.900 --> 00:45:23.020 align:middle line:84%
In particular, if it was true
before, it'll still be true.

00:45:23.020 --> 00:45:25.160 align:middle line:84%
If it was false before,
it'll still be false.

00:45:25.160 --> 00:45:26.170 align:middle line:90%
So here's a sentence.

00:45:26.170 --> 00:45:27.712 align:middle line:84%
"Professor Richards
is from Alabama."

00:45:27.712 --> 00:45:28.810 align:middle line:90%
That sentence is true.

00:45:28.810 --> 00:45:30.970 align:middle line:90%
I am from Alabama.

00:45:30.970 --> 00:45:35.680 align:middle line:84%
And if you were to switch out
the red expression for any

00:45:35.680 --> 00:45:37.570 align:middle line:84%
of the other red
expressions, the sentence

00:45:37.570 --> 00:45:38.600 align:middle line:90%
would still be true.

00:45:38.600 --> 00:45:40.810 align:middle line:84%
So to say that these all
mean the same thing--

00:45:40.810 --> 00:45:42.685 align:middle line:84%
to say that "Professor
Richards" and "Norvin"

00:45:42.685 --> 00:45:44.227 align:middle line:84%
and "that guy with
the beard" can all

00:45:44.227 --> 00:45:45.850 align:middle line:84%
refer to the same
person, is to say

00:45:45.850 --> 00:45:48.680 align:middle line:84%
that if the first
sentence is true,

00:45:48.680 --> 00:45:50.860 align:middle line:84%
then changing from one
of these to another one

00:45:50.860 --> 00:45:52.060 align:middle line:90%
is not going to change that.

00:45:52.060 --> 00:45:54.460 align:middle line:84%
If it's false, it's not
going to change that.

00:45:54.460 --> 00:45:55.060 align:middle line:90%
Sound right?

00:45:55.060 --> 00:45:59.760 align:middle line:90%


00:45:59.760 --> 00:46:01.772 align:middle line:84%
We have to be
careful, because it's

00:46:01.772 --> 00:46:04.230 align:middle line:84%
possible to get a little fancier
with the meanings of words

00:46:04.230 --> 00:46:06.600 align:middle line:90%
and phrases.

00:46:06.600 --> 00:46:08.700 align:middle line:84%
People sometimes
talk about what's

00:46:08.700 --> 00:46:13.680 align:middle line:84%
called the "intension" and
the "extension" of the meaning

00:46:13.680 --> 00:46:16.080 align:middle line:90%
of a phrase or a word.

00:46:16.080 --> 00:46:19.170 align:middle line:84%
Maybe the easiest way to think
about this, at least for me,

00:46:19.170 --> 00:46:21.090 align:middle line:90%
goes like this.

00:46:21.090 --> 00:46:24.450 align:middle line:84%
The intension of
a word, if you ask

00:46:24.450 --> 00:46:27.870 align:middle line:84%
what the intension of a word
is, what you're asking me for

00:46:27.870 --> 00:46:31.427 align:middle line:84%
is the procedure
that you will use.

00:46:31.427 --> 00:46:33.760 align:middle line:84%
I'm going to give you examples
of this in just a second.

00:46:33.760 --> 00:46:35.940 align:middle line:84%
But the intension
is the procedure

00:46:35.940 --> 00:46:39.060 align:middle line:84%
that you will use to determine
what that word or that phrase

00:46:39.060 --> 00:46:40.440 align:middle line:90%
refers to.

00:46:40.440 --> 00:46:42.870 align:middle line:84%
Whereas the extension
of a word or phrase

00:46:42.870 --> 00:46:46.110 align:middle line:84%
is the value, the
thing you will get

00:46:46.110 --> 00:46:48.730 align:middle line:84%
if you apply that procedure,
the value of that function.

00:46:48.730 --> 00:46:50.940 align:middle line:84%
Let me show you what the
heck I'm talking about.

00:46:50.940 --> 00:46:56.200 align:middle line:84%
So the phrase "the President
of the United States"

00:46:56.200 --> 00:46:58.870 align:middle line:84%
has an intension
and an extension.

00:46:58.870 --> 00:47:01.540 align:middle line:84%
The intension of that phrase,
"the President of the United

00:47:01.540 --> 00:47:04.720 align:middle line:84%
States" is the
procedure that you

00:47:04.720 --> 00:47:07.480 align:middle line:84%
should use to find out
who that phrase refers to.

00:47:07.480 --> 00:47:09.190 align:middle line:84%
So what you should
do is find out

00:47:09.190 --> 00:47:12.190 align:middle line:84%
who won the election,
who won the most

00:47:12.190 --> 00:47:14.900 align:middle line:90%
recent election for president.

00:47:14.900 --> 00:47:16.960 align:middle line:84%
That's the intention
of that phrase.

00:47:16.960 --> 00:47:20.080 align:middle line:84%
The extension of
that phrase is, well,

00:47:20.080 --> 00:47:22.660 align:middle line:84%
what you'll get if you
apply that procedure, which,

00:47:22.660 --> 00:47:26.060 align:middle line:90%
right now, is Joe Biden.

00:47:26.060 --> 00:47:29.450 align:middle line:84%
Where similarly, the phrase,
"the current temperature"

00:47:29.450 --> 00:47:31.310 align:middle line:84%
has an intension
and an extension.

00:47:31.310 --> 00:47:33.680 align:middle line:84%
The intension is
the procedure we

00:47:33.680 --> 00:47:35.960 align:middle line:84%
will use to find out what
the current temperature is,

00:47:35.960 --> 00:47:40.160 align:middle line:84%
like go look at the thermometer,
go on to weather.com,

00:47:40.160 --> 00:47:41.490 align:middle line:90%
or whatever.

00:47:41.490 --> 00:47:43.187 align:middle line:84%
And then it has an
extension, which

00:47:43.187 --> 00:47:44.270 align:middle line:90%
is, well, whatever it is--

00:47:44.270 --> 00:47:46.134 align:middle line:90%
45 degrees, let's say.

00:47:46.134 --> 00:47:50.220 align:middle line:90%


00:47:50.220 --> 00:47:55.800 align:middle line:84%
So I said before, when we
say that these phrases mean

00:47:55.800 --> 00:47:59.820 align:middle line:84%
the same thing, to say that
is to say that you could

00:47:59.820 --> 00:48:03.990 align:middle line:84%
take a sentence that's true
and switch one of the phrases

00:48:03.990 --> 00:48:06.430 align:middle line:84%
and still have a
sentence that's true.

00:48:06.430 --> 00:48:08.820 align:middle line:84%
We have to be careful,
though, about things

00:48:08.820 --> 00:48:10.870 align:middle line:90%
like intensions and extensions.

00:48:10.870 --> 00:48:14.460 align:middle line:84%
So when we say "The
temperature is 45 degrees,"

00:48:14.460 --> 00:48:18.390 align:middle line:84%
or "That guy with the beard
is Professor Richards,"

00:48:18.390 --> 00:48:20.160 align:middle line:84%
it's true that "that
guy with the beard"

00:48:20.160 --> 00:48:21.618 align:middle line:84%
and "Professor
Richards" are mostly

00:48:21.618 --> 00:48:23.940 align:middle line:90%
substitutable for each other.

00:48:23.940 --> 00:48:27.870 align:middle line:84%
But we have to be careful with
phrases like "the temperature"

00:48:27.870 --> 00:48:28.960 align:middle line:90%
and "45 degrees."

00:48:28.960 --> 00:48:31.980 align:middle line:84%
So the temperature has an
intension and an extension.

00:48:31.980 --> 00:48:34.830 align:middle line:84%
And its extension
might be 45 degrees,

00:48:34.830 --> 00:48:37.080 align:middle line:84%
but its intension
isn't necessarily,

00:48:37.080 --> 00:48:40.270 align:middle line:84%
which means that, for
example, even if it's true

00:48:40.270 --> 00:48:44.090 align:middle line:84%
that the temperature
is rising, it

00:48:44.090 --> 00:48:47.780 align:middle line:84%
doesn't follow that
45 degrees is rising,

00:48:47.780 --> 00:48:50.510 align:middle line:84%
even if it's true that the
temperature is 45 degrees

00:48:50.510 --> 00:48:52.490 align:middle line:84%
and that the
temperature is rising.

00:48:52.490 --> 00:48:56.000 align:middle line:84%
So here's a place where it seems
as though you can't substitute

00:48:56.000 --> 00:48:57.920 align:middle line:90%
in one phrase for another.

00:48:57.920 --> 00:49:03.800 align:middle line:84%
That's because rising is
interacting, specifically,

00:49:03.800 --> 00:49:04.760 align:middle line:90%
with the intension.

00:49:04.760 --> 00:49:06.843 align:middle line:84%
It's telling you something
about what you will see

00:49:06.843 --> 00:49:08.617 align:middle line:90%
if you look at the thermometer.

00:49:08.617 --> 00:49:10.700 align:middle line:84%
It's not telling you
something about the extension

00:49:10.700 --> 00:49:14.430 align:middle line:90%
of the phrase, the temperature.

00:49:14.430 --> 00:49:18.560 align:middle line:84%
So there are places where
you can confuse yourself

00:49:18.560 --> 00:49:20.180 align:middle line:90%
with intensions and extensions.

00:49:20.180 --> 00:49:23.270 align:middle line:90%


00:49:23.270 --> 00:49:27.430 align:middle line:84%
Let's extend the focus a little
bit and talk about sentences.

00:49:27.430 --> 00:49:30.023 align:middle line:84%
So that's the beginnings
of some things

00:49:30.023 --> 00:49:31.690 align:middle line:84%
we'll have to watch
out for when we look

00:49:31.690 --> 00:49:33.447 align:middle line:90%
at the meanings of phrases.

00:49:33.447 --> 00:49:35.530 align:middle line:84%
When we start talking
seriously about the meanings

00:49:35.530 --> 00:49:37.480 align:middle line:84%
of sentences, there
are some relations

00:49:37.480 --> 00:49:39.010 align:middle line:84%
between sentences
that it's going

00:49:39.010 --> 00:49:42.160 align:middle line:90%
to be useful to keep track of.

00:49:42.160 --> 00:49:44.130 align:middle line:90%
So one of them is entailment.

00:49:44.130 --> 00:49:48.630 align:middle line:84%
So we say that a sentence
A "entails" a sentence B,

00:49:48.630 --> 00:49:52.570 align:middle line:84%
if whenever A is
true, B must be true.

00:49:52.570 --> 00:49:56.190 align:middle line:84%
So if John killed the
ant, it has to be true

00:49:56.190 --> 00:49:57.420 align:middle line:90%
that the ant is dead.

00:49:57.420 --> 00:49:58.800 align:middle line:90%
That's what "kill" means.

00:49:58.800 --> 00:50:01.150 align:middle line:90%
Yes?

00:50:01.150 --> 00:50:02.775 align:middle line:84%
AUDIENCE: For this
temperature example,

00:50:02.775 --> 00:50:05.850 align:middle line:84%
we have the intensions
and the extensions,

00:50:05.850 --> 00:50:09.133 align:middle line:84%
and the intension
is not fixed, right?

00:50:09.133 --> 00:50:10.050 align:middle line:90%
NORVIN RICHARDS: Yeah.

00:50:10.050 --> 00:50:12.380 align:middle line:84%
AUDIENCE: It is 45 degrees
now, up to 50 degrees.

00:50:12.380 --> 00:50:12.980 align:middle line:90%
NORVIN RICHARDS: Sure.

00:50:12.980 --> 00:50:13.760 align:middle line:90%
AUDIENCE: Next day.

00:50:13.760 --> 00:50:14.677 align:middle line:90%
NORVIN RICHARDS: Yeah.

00:50:14.677 --> 00:50:18.410 align:middle line:84%
AUDIENCE: If we said "the winner
of the previous election,"

00:50:18.410 --> 00:50:24.290 align:middle line:84%
would that be always
substitutable for "Joe Biden"?

00:50:24.290 --> 00:50:26.940 align:middle line:84%
NORVIN RICHARDS: So--
oh, "the winner"--

00:50:26.940 --> 00:50:28.310 align:middle line:90%
so let's see now.

00:50:28.310 --> 00:50:32.330 align:middle line:84%
The phrase that-- let's
go back to Joe Biden.

00:50:32.330 --> 00:50:34.550 align:middle line:90%
Here he is.

00:50:34.550 --> 00:50:38.510 align:middle line:84%
It's not always going to
be true that "the winner

00:50:38.510 --> 00:50:41.720 align:middle line:84%
of the previous election"
is substitutable for "Joe

00:50:41.720 --> 00:50:48.440 align:middle line:84%
Biden," because I can say things
like "In 1945, the President

00:50:48.440 --> 00:50:51.560 align:middle line:84%
of the United States
declared war on Japan."

00:50:51.560 --> 00:50:53.870 align:middle line:84%
And that phrase,
and it's probably

00:50:53.870 --> 00:50:56.368 align:middle line:90%
the wrong year, that phrase--

00:50:56.368 --> 00:50:58.910 align:middle line:84%
even if that sentence is true,
I don't get to substitute that

00:50:58.910 --> 00:51:00.380 align:middle line:90%
in for "Joe Biden."

00:51:00.380 --> 00:51:04.850 align:middle line:84%
That's a case where it
refers to a different person,

00:51:04.850 --> 00:51:08.450 align:middle line:84%
to the person who won the
election previous to that.

00:51:08.450 --> 00:51:10.380 align:middle line:90%
Am I getting at your--

00:51:10.380 --> 00:51:12.540 align:middle line:84%
AUDIENCE: I feel like it
could stickier than that?

00:51:12.540 --> 00:51:14.370 align:middle line:90%
NORVIN RICHARDS: Probably.

00:51:14.370 --> 00:51:15.660 align:middle line:90%
This is day one of semantics.

00:51:15.660 --> 00:51:17.910 align:middle line:90%
But yes, you're right.

00:51:17.910 --> 00:51:22.408 align:middle line:84%
This is mainly meant to make
you wary about expressions

00:51:22.408 --> 00:51:23.700 align:middle line:90%
like "the meaning of a phrase."

00:51:23.700 --> 00:51:27.730 align:middle line:84%
So the meaning of a phrase
has more than one dimension.

00:51:27.730 --> 00:51:30.240 align:middle line:84%
You have to be careful about
what kinds of predicates

00:51:30.240 --> 00:51:31.335 align:middle line:90%
it's interacting with.

00:51:31.335 --> 00:51:33.990 align:middle line:90%


00:51:33.990 --> 00:51:37.740 align:middle line:84%
I mean, similarly,
if I say something

00:51:37.740 --> 00:51:41.340 align:middle line:84%
like, "The President
of the United States

00:51:41.340 --> 00:51:45.070 align:middle line:84%
has always been
male," that's true.

00:51:45.070 --> 00:51:47.880 align:middle line:84%
It's also true of Joe
Biden, as far as I know.

00:51:47.880 --> 00:51:50.345 align:middle line:84%
But it's-- those
aren't the same claim.

00:51:50.345 --> 00:52:02.570 align:middle line:90%


00:52:02.570 --> 00:52:05.990 align:middle line:90%
Rising-- ants-- dead ants.

00:52:05.990 --> 00:52:08.480 align:middle line:84%
Yes, so entailment--
one sentence

00:52:08.480 --> 00:52:10.220 align:middle line:84%
entails another
sentence, if whenever

00:52:10.220 --> 00:52:12.845 align:middle line:84%
the first sentence is true, the
second sentence has to be true.

00:52:12.845 --> 00:52:15.090 align:middle line:90%
That's what entailment means.

00:52:15.090 --> 00:52:17.780 align:middle line:84%
Similarly, to say
"Norman is Don's nephew,"

00:52:17.780 --> 00:52:20.990 align:middle line:84%
is to say that "Don
is Norman's uncle."

00:52:20.990 --> 00:52:23.750 align:middle line:84%
The first of those sentences
entails the second--

00:52:23.750 --> 00:52:27.350 align:middle line:90%


00:52:27.350 --> 00:52:27.975 align:middle line:90%
no, it doesn't.

00:52:27.975 --> 00:52:34.190 align:middle line:90%


00:52:34.190 --> 00:52:36.410 align:middle line:90%
Actually, this is a lie!

00:52:36.410 --> 00:52:39.140 align:middle line:84%
Neither of these sentences
entails the other,

00:52:39.140 --> 00:52:42.600 align:middle line:84%
unless you know that Norman
and Don are both male.

00:52:42.600 --> 00:52:44.730 align:middle line:84%
If you know that Norman
and Don are both male,

00:52:44.730 --> 00:52:47.410 align:middle line:90%
then they entail each other.

00:52:47.410 --> 00:52:50.120 align:middle line:90%
Otherwise, not.

00:52:50.120 --> 00:52:50.990 align:middle line:90%
I should fix that.

00:52:50.990 --> 00:52:55.040 align:middle line:90%


00:52:55.040 --> 00:52:56.510 align:middle line:90%
I think this one does work.

00:52:56.510 --> 00:53:00.500 align:middle line:84%
"If John is here, then Mary
is here, and John is here."

00:53:00.500 --> 00:53:02.690 align:middle line:84%
That first sentence,
that long sentence,

00:53:02.690 --> 00:53:05.520 align:middle line:90%
entails that "Mary is here."

00:53:05.520 --> 00:53:06.990 align:middle line:84%
So if that first
sentence is true,

00:53:06.990 --> 00:53:10.230 align:middle line:90%
the second sentence is true.

00:53:10.230 --> 00:53:13.160 align:middle line:84%
So these are
entailment relations.

00:53:13.160 --> 00:53:14.690 align:middle line:84%
Entailment relations
are not about

00:53:14.690 --> 00:53:16.980 align:middle line:84%
whether their sentences
actually are true.

00:53:16.980 --> 00:53:20.480 align:middle line:84%
So to say that sentence A
entails sentence B is just

00:53:20.480 --> 00:53:24.900 align:middle line:84%
to say, if you imagine that A
is true, then B is also true.

00:53:24.900 --> 00:53:30.530 align:middle line:84%
So "Joe Biden is a
bachelor" entails

00:53:30.530 --> 00:53:34.100 align:middle line:90%
that "Joe Biden is unmarried."

00:53:34.100 --> 00:53:37.020 align:middle line:84%
They're both false, but if
the first sentence were true,

00:53:37.020 --> 00:53:40.250 align:middle line:84%
the second sentence
would be true.

00:53:40.250 --> 00:53:44.320 align:middle line:84%
Does the second sentence
entail the first?

00:53:44.320 --> 00:53:45.790 align:middle line:90%
AUDIENCE: What is a bachelor?

00:53:45.790 --> 00:53:48.100 align:middle line:84%
NORVIN RICHARDS:
An unmarried man.

00:53:48.100 --> 00:53:51.270 align:middle line:84%
AUDIENCE: If Joe Biden
is a man and unmarried

00:53:51.270 --> 00:53:53.560 align:middle line:84%
he may be a bachelor,
but Joe Biden could

00:53:53.560 --> 00:53:54.820 align:middle line:90%
be a woman for all we know.

00:53:54.820 --> 00:53:55.600 align:middle line:90%
NORVIN RICHARDS: Right.

00:53:55.600 --> 00:53:56.100 align:middle line:90%
Yeah.

00:53:56.100 --> 00:53:57.400 align:middle line:90%
That's an example.

00:53:57.400 --> 00:53:58.990 align:middle line:84%
And actually, I
think I misspoke when

00:53:58.990 --> 00:54:00.823 align:middle line:84%
I said that a bachelor
was an unmarried man.

00:54:00.823 --> 00:54:03.370 align:middle line:84%
If his wife had died,
he would be unmarried,

00:54:03.370 --> 00:54:05.020 align:middle line:90%
I guess, but not a bachelor.

00:54:05.020 --> 00:54:07.480 align:middle line:84%
I think a bachelor is someone
who has never married.

00:54:07.480 --> 00:54:08.230 align:middle line:90%
Yeah.

00:54:08.230 --> 00:54:09.380 align:middle line:90%
Also false.

00:54:09.380 --> 00:54:13.080 align:middle line:90%
So he's actually married.

00:54:13.080 --> 00:54:15.870 align:middle line:84%
So entailment relations are
not necessarily symmetrical.

00:54:15.870 --> 00:54:19.440 align:middle line:84%
It's possible for A to entail
B, and for B to not entail A.

00:54:19.440 --> 00:54:22.170 align:middle line:84%
And entailment
relations are about what

00:54:22.170 --> 00:54:24.390 align:middle line:84%
life would be like if the
first sentence were true.

00:54:24.390 --> 00:54:28.400 align:middle line:84%
It's not about
whether it is true.

00:54:28.400 --> 00:54:31.040 align:middle line:84%
And we say that A and B are
in an equivalence relation,

00:54:31.040 --> 00:54:33.343 align:middle line:90%
if they entail each other.

00:54:33.343 --> 00:54:35.010 align:middle line:84%
We can also say that
they're synonymous.

00:54:35.010 --> 00:54:38.367 align:middle line:84%
So "Mary ate the bagel" and
"The bagel was eaten by Mary."

00:54:38.367 --> 00:54:39.950 align:middle line:84%
The first sentence
entails the second.

00:54:39.950 --> 00:54:41.510 align:middle line:84%
And the second
entails the first.

00:54:41.510 --> 00:54:43.430 align:middle line:90%
These are equivalent sentences.

00:54:43.430 --> 00:54:45.813 align:middle line:90%


00:54:45.813 --> 00:54:48.230 align:middle line:84%
And we say that they contradict
each other if each entails

00:54:48.230 --> 00:54:49.740 align:middle line:90%
that the other is false.

00:54:49.740 --> 00:54:56.140 align:middle line:84%
So as long as "Noam" always
refers to the same person,

00:54:56.140 --> 00:55:00.040 align:middle line:84%
if Noam is here, that
entails that it is not true

00:55:00.040 --> 00:55:02.490 align:middle line:90%
that Noam is not here.

00:55:02.490 --> 00:55:06.350 align:middle line:84%
And if Noam is not here, that
entails that it is not true

00:55:06.350 --> 00:55:09.980 align:middle line:90%
that Naom is here.

00:55:09.980 --> 00:55:11.690 align:middle line:84%
I'm waiting for
somebody to work hard

00:55:11.690 --> 00:55:13.880 align:middle line:84%
on the meaning of
"here," which I

00:55:13.880 --> 00:55:15.830 align:middle line:84%
guess we could
suppose he's halfway

00:55:15.830 --> 00:55:17.270 align:middle line:90%
into the room or something.

00:55:17.270 --> 00:55:20.030 align:middle line:84%
But forget about that
stuff for a second.

00:55:20.030 --> 00:55:25.500 align:middle line:84%
Let's pretend that people are
either here or there, not here.

00:55:25.500 --> 00:55:27.560 align:middle line:84%
So we say that A and B
contradict each other,

00:55:27.560 --> 00:55:33.300 align:middle line:84%
if they each entail the
falsehood of the other.

00:55:33.300 --> 00:55:33.800 align:middle line:90%
Yeah?

00:55:33.800 --> 00:55:37.590 align:middle line:84%
AUDIENCE: I assume that
it's possible for A--

00:55:37.590 --> 00:55:41.210 align:middle line:84%
is it possible for A to
contradict B, but B not

00:55:41.210 --> 00:55:43.710 align:middle line:90%
contradict A as false?

00:55:43.710 --> 00:55:46.110 align:middle line:84%
NORVIN RICHARDS: So for A
to entail that B is false,

00:55:46.110 --> 00:55:48.540 align:middle line:84%
but for B to not
entail that A is false?

00:55:48.540 --> 00:55:50.040 align:middle line:90%
Yeah.

00:55:50.040 --> 00:55:57.200 align:middle line:84%
So if you go back to
this slide, because I

00:55:57.200 --> 00:56:00.230 align:middle line:84%
feel as though if we arrange
things, correctly-- yeah.

00:56:00.230 --> 00:56:05.870 align:middle line:84%
So "If Joe Biden
is a bachelor"...

00:56:05.870 --> 00:56:09.720 align:middle line:90%


00:56:09.720 --> 00:56:11.195 align:middle line:90%
nope.

00:56:11.195 --> 00:56:12.195 align:middle line:90%
What's a better example?

00:56:12.195 --> 00:56:15.540 align:middle line:90%


00:56:15.540 --> 00:56:21.650 align:middle line:84%
So if-- so we're looking for a
case of a subset relation where

00:56:21.650 --> 00:56:23.488 align:middle line:90%
all--

00:56:23.488 --> 00:56:25.280 align:middle line:84%
AUDIENCE: "If the
rectangle is a square..."

00:56:25.280 --> 00:56:26.510 align:middle line:84%
NORVIN RICHARDS: Is a
rectangle and square-- yeah,

00:56:26.510 --> 00:56:27.560 align:middle line:90%
that's better, isn't it?

00:56:27.560 --> 00:56:28.220 align:middle line:90%
Yeah, go ahead.

00:56:28.220 --> 00:56:29.928 align:middle line:84%
Give us the rectangle
and square example.

00:56:29.928 --> 00:56:33.202 align:middle line:90%


00:56:33.202 --> 00:56:37.220 align:middle line:84%
AUDIENCE: If you say that
something is not a rectangle,

00:56:37.220 --> 00:56:39.440 align:middle line:84%
that doesn't necessarily
mean that it isn't square.

00:56:39.440 --> 00:56:39.710 align:middle line:90%
NORVIN RICHARDS: Yeah.

00:56:39.710 --> 00:56:41.330 align:middle line:84%
AUDIENCE: If it
isn't a square, that

00:56:41.330 --> 00:56:42.830 align:middle line:84%
means it's definitely
not rectangle.

00:56:42.830 --> 00:56:43.130 align:middle line:90%
NORVIN RICHARDS: Yeah.

00:56:43.130 --> 00:56:43.630 align:middle line:90%
OK.

00:56:43.630 --> 00:56:44.330 align:middle line:90%
That's--

00:56:44.330 --> 00:56:44.840 align:middle line:90%
AUDIENCE: Other way around.

00:56:44.840 --> 00:56:45.830 align:middle line:84%
NORVIN RICHARDS: OK,
other way around.

00:56:45.830 --> 00:56:46.430 align:middle line:90%
Good.

00:56:46.430 --> 00:56:46.940 align:middle line:90%
Yes.

00:56:46.940 --> 00:56:48.890 align:middle line:84%
AUDIENCE: I think
that the logic rules

00:56:48.890 --> 00:56:53.960 align:middle line:84%
are like if you have something
like, "if p, then q," then

00:56:53.960 --> 00:56:57.530 align:middle line:84%
it follows that "not
q implies not p."

00:56:57.530 --> 00:57:01.425 align:middle line:84%
But you have no idea
whether q implies p or not.

00:57:01.425 --> 00:57:02.300 align:middle line:90%
NORVIN RICHARDS: Yes.

00:57:02.300 --> 00:57:02.810 align:middle line:90%
Yes.

00:57:02.810 --> 00:57:03.310 align:middle line:90%
Yes.

00:57:03.310 --> 00:57:04.130 align:middle line:90%
Yes, that's true.

00:57:04.130 --> 00:57:08.600 align:middle line:84%
So that is what people
say about [INAUDIBLE]..

00:57:08.600 --> 00:57:11.450 align:middle line:90%
Yeah, you're right.

00:57:11.450 --> 00:57:12.440 align:middle line:90%
Yes.

00:57:12.440 --> 00:57:15.370 align:middle line:90%
OK, good.

00:57:15.370 --> 00:57:17.200 align:middle line:84%
I'm going to move
away from entailment

00:57:17.200 --> 00:57:20.470 align:middle line:84%
and start talking
about presupposition.

00:57:20.470 --> 00:57:23.095 align:middle line:84%
Presupposition is another kind
of relation between sentences.

00:57:23.095 --> 00:57:25.720 align:middle line:84%
Again, it doesn't matter whether
the sentences are true or not,

00:57:25.720 --> 00:57:27.430 align:middle line:84%
you're being invited
to imagine what life

00:57:27.430 --> 00:57:29.960 align:middle line:90%
would be if they were true.

00:57:29.960 --> 00:57:35.370 align:middle line:84%
So a sentence like "The
present King of France is bald"

00:57:35.370 --> 00:57:40.750 align:middle line:84%
presupposes that there is
a present King of France.

00:57:40.750 --> 00:57:44.685 align:middle line:84%
So this is different
from entailment, in ways

00:57:44.685 --> 00:57:46.560 align:middle line:84%
that I think I illustrate
on the next slides.

00:57:46.560 --> 00:57:47.568 align:middle line:90%
Yes.

00:57:47.568 --> 00:57:49.110 align:middle line:84%
Crucially different
from entitlement.

00:57:49.110 --> 00:57:53.760 align:middle line:84%
So if I say "The present
King of France is bald,"

00:57:53.760 --> 00:57:57.460 align:middle line:84%
or if I say "The present
King of France is not bald,"

00:57:57.460 --> 00:58:00.250 align:middle line:84%
or if I say-- if I ask you,
"Is the present King of France

00:58:00.250 --> 00:58:04.780 align:middle line:84%
bald," all of those have this
relation of presupposition

00:58:04.780 --> 00:58:07.060 align:middle line:84%
to there is a present
King of France.

00:58:07.060 --> 00:58:09.580 align:middle line:84%
That is-- we can stick
to the first two--

00:58:09.580 --> 00:58:11.500 align:middle line:84%
If I say "The present
King of France is bald,"

00:58:11.500 --> 00:58:13.828 align:middle line:84%
or if I say "The present
King of France is not bald,"

00:58:13.828 --> 00:58:15.370 align:middle line:84%
for either of those
to be true, there

00:58:15.370 --> 00:58:18.650 align:middle line:84%
has to be a present
King of France.

00:58:18.650 --> 00:58:21.020 align:middle line:84%
That's different
from entailment.

00:58:21.020 --> 00:58:24.320 align:middle line:84%
So for entailment,
when we did things

00:58:24.320 --> 00:58:33.800 align:middle line:84%
like "John killed an
ant," and "The ant died,"

00:58:33.800 --> 00:58:35.450 align:middle line:90%
that's an entailment relation.

00:58:35.450 --> 00:58:39.110 align:middle line:84%
If I say "John didn't
kill an ant," well,

00:58:39.110 --> 00:58:41.120 align:middle line:84%
if John didn't kill
an ant, then the ant

00:58:41.120 --> 00:58:43.400 align:middle line:84%
didn't die, at least,
not necessarily.

00:58:43.400 --> 00:58:44.330 align:middle line:90%
It might be alive.

00:58:44.330 --> 00:58:48.272 align:middle line:84%
So to say that John killed an
ant entails that the ant died.

00:58:48.272 --> 00:58:49.730 align:middle line:84%
And if I negate
the first sentence,

00:58:49.730 --> 00:58:52.940 align:middle line:84%
I no longer entail
what I used to entail.

00:58:52.940 --> 00:58:56.220 align:middle line:84%
But if you negate a sentence
that presupposes something,

00:58:56.220 --> 00:58:57.180 align:middle line:90%
so if you add not--

00:58:57.180 --> 00:58:58.805 align:middle line:84%
if you go from "The
present King France

00:58:58.805 --> 00:59:00.890 align:middle line:84%
is bald" to "The present
King of France is not bald,"

00:59:00.890 --> 00:59:03.050 align:middle line:84%
you still have the
preposition-- the preposition--

00:59:03.050 --> 00:59:05.120 align:middle line:84%
the presuppositions
that you used to have.

00:59:05.120 --> 00:59:06.800 align:middle line:84%
You still presuppose--
it still has

00:59:06.800 --> 00:59:09.980 align:middle line:84%
to be true that there is
a present King of France.

00:59:09.980 --> 00:59:11.198 align:middle line:90%
Yes?

00:59:11.198 --> 00:59:13.490 align:middle line:84%
AUDIENCE: Let's say there is
no current King of France.

00:59:13.490 --> 00:59:15.013 align:middle line:84%
NORVIN RICHARDS:
Yes, which is true,

00:59:15.013 --> 00:59:16.430 align:middle line:84%
by the way, if
anybody is learning

00:59:16.430 --> 00:59:19.413 align:middle line:90%
about the world from this class.

00:59:19.413 --> 00:59:21.140 align:middle line:90%
AUDIENCE: [INAUDIBLE]

00:59:21.140 --> 00:59:22.107 align:middle line:90%


00:59:22.107 --> 00:59:22.940 align:middle line:90%
NORVIN RICHARDS: Oh.

00:59:22.940 --> 00:59:25.820 align:middle line:84%
So if there is no present
King of France, ah.

00:59:25.820 --> 00:59:28.820 align:middle line:84%
If there is no present
King of France--

00:59:28.820 --> 00:59:37.820 align:middle line:84%
good point-- their truth
value becomes difficult.

00:59:37.820 --> 00:59:40.560 align:middle line:84%
This is sometimes called
presupposition failure.

00:59:40.560 --> 00:59:45.633 align:middle line:84%
So if I say a sentence that
presupposes something false--

00:59:45.633 --> 00:59:47.300 align:middle line:84%
we're actually going
to talk about this,

00:59:47.300 --> 00:59:49.400 align:middle line:90%
I think, on the next slide--

00:59:49.400 --> 00:59:51.990 align:middle line:84%
your cognitive state
becomes difficult.

00:59:51.990 --> 00:59:57.370 align:middle line:90%
So this is a classic of--

00:59:57.370 --> 01:00:00.910 align:middle line:84%
it's used in politics
and also in comedy.

01:00:00.910 --> 01:00:02.897 align:middle line:84%
Unfortunate example
from the Marx brothers--

01:00:02.897 --> 01:00:04.480 align:middle line:84%
the Marx brothers
asked at some point,

01:00:04.480 --> 01:00:06.022 align:middle line:84%
"Have you stopped
beating your wife?"

01:00:06.022 --> 01:00:08.410 align:middle line:90%


01:00:08.410 --> 01:00:12.730 align:middle line:84%
The point of the question is
that there's no good answer.

01:00:12.730 --> 01:00:18.370 align:middle line:84%
If you say, yes or no, you're
accepting the presupposition

01:00:18.370 --> 01:00:19.720 align:middle line:90%
that you used to beat your wife.

01:00:19.720 --> 01:00:22.990 align:middle line:84%
That's what this
question does, right?

01:00:22.990 --> 01:00:26.140 align:middle line:84%
The corresponding statement
has the same presupposition.

01:00:26.140 --> 01:00:33.580 align:middle line:84%
So "He stopped
smoking" presupposes

01:00:33.580 --> 01:00:36.190 align:middle line:90%
that he used to smoke.

01:00:36.190 --> 01:00:42.040 align:middle line:84%
And that means that
if I presuppose

01:00:42.040 --> 01:00:44.020 align:middle line:84%
something which is
false, so if you ask me,

01:00:44.020 --> 01:00:46.655 align:middle line:84%
have you stopped smoking,
you're presupposing

01:00:46.655 --> 01:00:47.530 align:middle line:90%
that I used to smoke.

01:00:47.530 --> 01:00:49.210 align:middle line:84%
And if that's false,
it's not enough

01:00:49.210 --> 01:00:50.920 align:middle line:90%
for me to either say yes or no.

01:00:50.920 --> 01:00:53.530 align:middle line:90%
I inherit your presupposition.

01:00:53.530 --> 01:00:55.010 align:middle line:90%
I have to say something.

01:00:55.010 --> 01:00:57.220 align:middle line:84%
I have to say something
like, "Wait, I never

01:00:57.220 --> 01:00:59.020 align:middle line:90%
smoked in the first place.

01:00:59.020 --> 01:01:01.217 align:middle line:90%
Your presupposition is false"--

01:01:01.217 --> 01:01:02.050 align:middle line:90%
something like that.

01:01:02.050 --> 01:01:06.400 align:middle line:84%
There's-- in the
literature on Zen Buddhism,

01:01:06.400 --> 01:01:10.150 align:middle line:84%
which I'm not an expert on,
there's a kind of exchange

01:01:10.150 --> 01:01:13.330 align:middle line:84%
between teachers and students
that happens a lot where

01:01:13.330 --> 01:01:15.727 align:middle line:84%
a teacher will ask a
student the question.

01:01:15.727 --> 01:01:17.560 align:middle line:84%
The teacher-- the student
will ask something

01:01:17.560 --> 01:01:21.190 align:middle line:84%
like, "Does the dog
have the Buddha nature?"

01:01:21.190 --> 01:01:22.190 align:middle line:90%
or something like that.

01:01:22.190 --> 01:01:24.910 align:middle line:84%
And the teacher,
by way of answer,

01:01:24.910 --> 01:01:27.550 align:middle line:84%
will haul off and
slap the student,

01:01:27.550 --> 01:01:31.870 align:middle line:84%
or pour cold water on the
student, or shout "mu!"

01:01:31.870 --> 01:01:34.150 align:middle line:90%
and the student is enlightened.

01:01:34.150 --> 01:01:36.970 align:middle line:84%
This is a kind of story you
see a lot in the Zen Buddhism

01:01:36.970 --> 01:01:38.407 align:middle line:90%
literature.

01:01:38.407 --> 01:01:40.240 align:middle line:84%
I don't know a whole
lot about Zen Buddhism,

01:01:40.240 --> 01:01:42.310 align:middle line:84%
but I've always
wondered whether "mu!"

01:01:42.310 --> 01:01:45.860 align:middle line:84%
means you are making a
false presupposition.

01:01:45.860 --> 01:01:47.860 align:middle line:84%
Because it's an interesting
fact about languages

01:01:47.860 --> 01:01:49.900 align:middle line:84%
that we have words
like "yes" or "no"

01:01:49.900 --> 01:01:54.070 align:middle line:84%
that say things like
"The statement that

01:01:54.070 --> 01:01:56.770 align:middle line:84%
underlies your question
is true or false."

01:01:56.770 --> 01:01:58.390 align:middle line:90%
If I ask you "Is it raining?"

01:01:58.390 --> 01:02:01.510 align:middle line:84%
you're asking me to evaluate
the truth or falsehood of "It

01:02:01.510 --> 01:02:02.320 align:middle line:90%
is raining?"

01:02:02.320 --> 01:02:04.540 align:middle line:84%
And I'm supposed to
say either "Yes, it's

01:02:04.540 --> 01:02:06.040 align:middle line:84%
true that it is
raining," or "No,

01:02:06.040 --> 01:02:08.010 align:middle line:90%
it's false that it's raining."

01:02:08.010 --> 01:02:10.502 align:middle line:84%
You ask me "Have you
stopped smoking?"

01:02:10.502 --> 01:02:12.460 align:middle line:84%
the statement that you're
asking me to evaluate

01:02:12.460 --> 01:02:14.060 align:middle line:90%
is "You have stopped smoking."

01:02:14.060 --> 01:02:17.770 align:middle line:84%
And I have "yes" and "no"
as my standard options.

01:02:17.770 --> 01:02:19.810 align:middle line:84%
But I don't have an
option that says there

01:02:19.810 --> 01:02:21.430 align:middle line:90%
is a false presupposition.

01:02:21.430 --> 01:02:23.492 align:middle line:90%
And most languages don't.

01:02:23.492 --> 01:02:25.450 align:middle line:84%
It's not common for
languages to have something

01:02:25.450 --> 01:02:29.560 align:middle line:84%
to say that means you have
made a false presupposition.

01:02:29.560 --> 01:02:32.980 align:middle line:84%
And so it's interesting
that language work this way.

01:02:32.980 --> 01:02:36.160 align:middle line:84%
AUDIENCE: That's interesting,
because I think in math,

01:02:36.160 --> 01:02:40.780 align:middle line:90%
like in formal logic, if the--

01:02:40.780 --> 01:02:43.550 align:middle line:84%
if this, then that, then
if this is just false,

01:02:43.550 --> 01:02:47.410 align:middle line:84%
then it's the whole thing
is true just by default.

01:02:47.410 --> 01:02:49.180 align:middle line:84%
And the thing I'm
wondering is like,

01:02:49.180 --> 01:02:51.910 align:middle line:84%
does this mean that
when you ask a question,

01:02:51.910 --> 01:02:53.798 align:middle line:84%
you aren't actually
asking, like "Have

01:02:53.798 --> 01:02:55.090 align:middle line:90%
you stopped beating your wife?"

01:02:55.090 --> 01:02:58.120 align:middle line:84%
You're asking, like, "Are
all the presuppositions

01:02:58.120 --> 01:03:01.233 align:middle line:84%
in this true, and
is this also true?"

01:03:01.233 --> 01:03:02.650 align:middle line:84%
NORVIN RICHARDS:
So that's just it

01:03:02.650 --> 01:03:05.050 align:middle line:84%
though, I'm not asking you
whether the presuppositions are

01:03:05.050 --> 01:03:05.550 align:middle line:90%
true.

01:03:05.550 --> 01:03:09.340 align:middle line:84%
That's why if I ask
you this question--

01:03:09.340 --> 01:03:12.520 align:middle line:84%
let's change it to "Have
you stopped smoking?"--

01:03:12.520 --> 01:03:14.590 align:middle line:84%
If I ask you "Have
you stopped smoking?"

01:03:14.590 --> 01:03:19.060 align:middle line:84%
you can't say no
and have that mean--

01:03:19.060 --> 01:03:22.078 align:middle line:84%
and have me interpret that
as meaning you never smoked

01:03:22.078 --> 01:03:22.870 align:middle line:90%
in the first place.

01:03:22.870 --> 01:03:28.630 align:middle line:84%
You're absolutely right
that if you were--

01:03:28.630 --> 01:03:31.180 align:middle line:84%
this is the kind of thing
that robots and aliens

01:03:31.180 --> 01:03:33.880 align:middle line:84%
get wrong, standardly, in
science fiction stories,

01:03:33.880 --> 01:03:38.320 align:middle line:84%
that false presuppositions
lead to ungrammaticality,

01:03:38.320 --> 01:03:39.610 align:middle line:90%
or something like that.

01:03:39.610 --> 01:03:41.740 align:middle line:84%
There's something
weird going on.

01:03:41.740 --> 01:03:44.330 align:middle line:84%
Did I get it, the
answer to your question?

01:03:44.330 --> 01:03:46.060 align:middle line:84%
So presupposition
failure, this case

01:03:46.060 --> 01:03:49.120 align:middle line:84%
where I say something that
has a false presupposition, it

01:03:49.120 --> 01:03:52.930 align:middle line:84%
doesn't make the sentence false,
it makes the sentence weird.

01:03:52.930 --> 01:03:57.550 align:middle line:84%
Yeah, that's the expression,
or the feeling that we have.

01:03:57.550 --> 01:03:59.330 align:middle line:84%
I skipped some slides
to get to this one.

01:03:59.330 --> 01:04:00.330 align:middle line:90%
So let me go back a bit.

01:04:00.330 --> 01:04:04.010 align:middle line:90%


01:04:04.010 --> 01:04:07.670 align:middle line:84%
So "The present King
of France is bald"

01:04:07.670 --> 01:04:09.900 align:middle line:84%
presupposes that there is
a present King of France.

01:04:09.900 --> 01:04:11.650 align:middle line:84%
It entails that the
present King of France

01:04:11.650 --> 01:04:15.180 align:middle line:84%
has no hair, because
that's what it means.

01:04:15.180 --> 01:04:20.330 align:middle line:84%
And so, if I say "The present
King of France is not bald,"

01:04:20.330 --> 01:04:22.670 align:middle line:84%
that no longer entails
that he has no hair.

01:04:22.670 --> 01:04:24.590 align:middle line:84%
That's the difference
between entailments

01:04:24.590 --> 01:04:25.700 align:middle line:90%
and presuppositions.

01:04:25.700 --> 01:04:29.120 align:middle line:84%
Entailments don't
survive negation.

01:04:29.120 --> 01:04:31.700 align:middle line:84%
If you add "not," then
you lose the entailments

01:04:31.700 --> 01:04:32.900 align:middle line:90%
that you used to have.

01:04:32.900 --> 01:04:34.820 align:middle line:84%
But you still have
the presuppositions

01:04:34.820 --> 01:04:36.500 align:middle line:90%
that you used to have.

01:04:36.500 --> 01:04:37.460 align:middle line:90%
Yeah?

01:04:37.460 --> 01:04:41.440 align:middle line:84%
AUDIENCE: Did that say
that entails the opposite?

01:04:41.440 --> 01:04:42.690 align:middle line:90%
NORVIN RICHARDS: Say it again.

01:04:42.690 --> 01:04:44.260 align:middle line:84%
AUDIENCE: Did you then
just say that entailment

01:04:44.260 --> 01:04:44.927 align:middle line:90%
is the opposite?

01:04:44.927 --> 01:04:47.070 align:middle line:84%
Like it entails
that he has hair?

01:04:47.070 --> 01:04:48.945 align:middle line:84%
NORVIN RICHARDS: Oh, in
this particular case?

01:04:48.945 --> 01:04:52.170 align:middle line:90%


01:04:52.170 --> 01:04:53.640 align:middle line:90%
Yes.

01:04:53.640 --> 01:05:00.580 align:middle line:84%
I think we're going to find
examples where A entails B,

01:05:00.580 --> 01:05:05.860 align:middle line:84%
But not A does not entail
not B. So yeah, if something

01:05:05.860 --> 01:05:09.698 align:middle line:90%
is a rectangle, then it-- no.

01:05:09.698 --> 01:05:11.740 align:middle line:84%
If something is a square,
then it is a rectangle.

01:05:11.740 --> 01:05:13.540 align:middle line:90%
Did I get that right?

01:05:13.540 --> 01:05:16.750 align:middle line:84%
But if something
is not a square,

01:05:16.750 --> 01:05:18.920 align:middle line:84%
that doesn't mean that
it's not a rectangle.

01:05:18.920 --> 01:05:22.690 align:middle line:84%
So yes, that's a case where A
entails B, but not A doesn't

01:05:22.690 --> 01:05:25.910 align:middle line:90%
entail not B. Yeah, good.

01:05:25.910 --> 01:05:26.410 align:middle line:90%
Phew.

01:05:26.410 --> 01:05:36.120 align:middle line:90%


01:05:36.120 --> 01:05:38.680 align:middle line:84%
And then presuppositions--
yeah, presuppositions are funny.

01:05:38.680 --> 01:05:43.500 align:middle line:84%
So if depending on
the presupposition--

01:05:43.500 --> 01:05:46.050 align:middle line:84%
so if we hear scratching
at the door, I could say,

01:05:46.050 --> 01:05:49.425 align:middle line:84%
"The cat is at the
door," out of the blue.

01:05:49.425 --> 01:05:51.610 align:middle line:84%
But there's a
presupposition there,

01:05:51.610 --> 01:05:53.220 align:middle line:90%
which is that "There is a cat."

01:05:53.220 --> 01:05:55.740 align:middle line:84%
It's kind of like "The present
King of France is bald"

01:05:55.740 --> 01:05:59.070 align:middle line:84%
presupposes that there
is a King of France.

01:05:59.070 --> 01:06:02.460 align:middle line:84%
There's a process, what's called
accommodating presuppositions,

01:06:02.460 --> 01:06:04.650 align:middle line:84%
and this is why
presuppositions are so handy

01:06:04.650 --> 01:06:06.570 align:middle line:90%
in comedy and in politics.

01:06:06.570 --> 01:06:11.880 align:middle line:84%
You make-- you say things that
presuppose things, and then

01:06:11.880 --> 01:06:14.130 align:middle line:84%
people accommodate your
presuppositions sometimes

01:06:14.130 --> 01:06:15.310 align:middle line:90%
without realizing it.

01:06:15.310 --> 01:06:18.253 align:middle line:84%
So if I, out of the blue,
say, "The cat is at the door,"

01:06:18.253 --> 01:06:19.920 align:middle line:84%
you didn't know before
that I had a cat.

01:06:19.920 --> 01:06:21.253 align:middle line:90%
I hadn't mentioned a cat before.

01:06:21.253 --> 01:06:25.200 align:middle line:84%
But my sentence presupposes
that there is a cat.

01:06:25.200 --> 01:06:27.960 align:middle line:84%
And so you learn from my
sentence that there is a cat

01:06:27.960 --> 01:06:32.580 align:middle line:84%
and you add that to
your knowledge base.

01:06:32.580 --> 01:06:36.300 align:middle line:84%
If I were to instead say,
"The giraffe is at the door,"

01:06:36.300 --> 01:06:39.390 align:middle line:84%
you'd probably be less
peaceful about it.

01:06:39.390 --> 01:06:42.120 align:middle line:84%
So if I make you
presuppose something

01:06:42.120 --> 01:06:45.570 align:middle line:84%
that's easy to presuppose,
you do it without noticing it.

01:06:45.570 --> 01:06:51.180 align:middle line:84%
It's not like you consciously
go through a state of,

01:06:51.180 --> 01:06:52.660 align:middle line:84%
if I say "The cat
is at the door,"

01:06:52.660 --> 01:06:54.160 align:middle line:84%
it's not like you
think to yourself,

01:06:54.160 --> 01:06:57.498 align:middle line:90%
"Ah, he must have a cat.

01:06:57.498 --> 01:06:59.790 align:middle line:84%
Write down in your mental
notebook he must have a cat."

01:06:59.790 --> 01:07:02.040 align:middle line:84%
Maybe that's happening
on some very fast level,

01:07:02.040 --> 01:07:04.075 align:middle line:90%
but you're not conscious of it.

01:07:04.075 --> 01:07:06.450 align:middle line:84%
On the other hand, if I say
"The giraffe is at the door,"

01:07:06.450 --> 01:07:08.580 align:middle line:84%
then you have the
sensation of having

01:07:08.580 --> 01:07:10.170 align:middle line:90%
learned something significant.

01:07:10.170 --> 01:07:14.420 align:middle line:90%


01:07:14.420 --> 01:07:17.810 align:middle line:84%
Right, and similarly, "I regret
having been born in 1857."

01:07:17.810 --> 01:07:20.150 align:middle line:84%
If I say that and
somebody objects,

01:07:20.150 --> 01:07:23.840 align:middle line:84%
I say, OK, "I don't regret
having been born in 1857."

01:07:23.840 --> 01:07:27.320 align:middle line:84%
Either of these presupposes
that I was born in 1857.

01:07:27.320 --> 01:07:30.297 align:middle line:84%
And in order for
somebody to object,

01:07:30.297 --> 01:07:31.880 align:middle line:84%
they have to say
"No, wait, surely you

01:07:31.880 --> 01:07:35.400 align:middle line:90%
were not born in 1857."

01:07:35.400 --> 01:07:37.890 align:middle line:84%
So that's entailments
and presuppositions.

01:07:37.890 --> 01:07:40.710 align:middle line:84%
There are also implicatures,
which I think we have time for.

01:07:40.710 --> 01:07:42.660 align:middle line:84%
The implicatures
are things that you

01:07:42.660 --> 01:07:45.060 align:middle line:84%
might tend to infer
from hearing a sentence

01:07:45.060 --> 01:07:47.260 align:middle line:84%
but that might not
be true at all.

01:07:47.260 --> 01:07:51.300 align:middle line:84%
So, for example, if I
ask you a question, like

01:07:51.300 --> 01:07:53.832 align:middle line:90%
"Can you open the window?"

01:07:53.832 --> 01:07:55.290 align:middle line:84%
there's an implicature
that I would

01:07:55.290 --> 01:07:58.100 align:middle line:90%
like you to open the window.

01:07:58.100 --> 01:08:01.040 align:middle line:84%
It's not entailed
or presupposed.

01:08:01.040 --> 01:08:03.080 align:middle line:84%
But you, when you
hear me say that,

01:08:03.080 --> 01:08:04.910 align:middle line:84%
you're likely to
conclude that I wish

01:08:04.910 --> 01:08:06.230 align:middle line:90%
that you would open the window.

01:08:06.230 --> 01:08:11.090 align:middle line:84%
Or if I ask you,
"Where's the salt?"

01:08:11.090 --> 01:08:14.390 align:middle line:84%
there is a minimally
cooperative answer to that

01:08:14.390 --> 01:08:17.029 align:middle line:90%
in which you point, you know.

01:08:17.029 --> 01:08:20.569 align:middle line:84%
But what I really want you
to do is give me some salt.

01:08:20.569 --> 01:08:25.130 align:middle line:84%
Or if I say I am 21,
there is an implication

01:08:25.130 --> 01:08:28.175 align:middle line:84%
that I am 21, that, I'm,
in fact, exactly 21.

01:08:28.175 --> 01:08:30.770 align:middle line:90%
21 is my age.

01:08:30.770 --> 01:08:34.160 align:middle line:84%
Or if I say "Mary ate
some of the cookies,"

01:08:34.160 --> 01:08:38.479 align:middle line:84%
there's an implicature that
there are some cookies left.

01:08:38.479 --> 01:08:40.910 align:middle line:84%
But all of these
things could be false,

01:08:40.910 --> 01:08:42.720 align:middle line:90%
under the right circumstances.

01:08:42.720 --> 01:08:46.010 align:middle line:84%
So if I ask you, "Can
you open the window?"

01:08:46.010 --> 01:08:48.680 align:middle line:84%
probably what I-- under most
circumstances, what I want

01:08:48.680 --> 01:08:50.090 align:middle line:90%
you to do is open the window.

01:08:50.090 --> 01:08:52.550 align:middle line:84%
But maybe what I'm
doing is making

01:08:52.550 --> 01:08:54.630 align:middle line:84%
a study of physical
fitness in MIT students

01:08:54.630 --> 01:08:56.005 align:middle line:84%
and I want to know
whether you're

01:08:56.005 --> 01:08:57.740 align:middle line:90%
capable of opening the window.

01:08:57.740 --> 01:09:01.090 align:middle line:84%
Or "Where's the salt?"--
maybe, in most circumstances,

01:09:01.090 --> 01:09:02.840 align:middle line:84%
what I want you to do
is hand me the salt,

01:09:02.840 --> 01:09:05.330 align:middle line:84%
but maybe I'm making
a map of the kitchen.

01:09:05.330 --> 01:09:07.939 align:middle line:84%
Or there are circumstances
in which I could say,

01:09:07.939 --> 01:09:10.970 align:middle line:84%
I, personally, could
say, I'm 21, namely,

01:09:10.970 --> 01:09:13.100 align:middle line:84%
when I'm trying to go
into a bar or something,

01:09:13.100 --> 01:09:16.370 align:middle line:84%
and they tell you "You can't
come in here unless you're 21."

01:09:16.370 --> 01:09:19.740 align:middle line:90%
I could say "I'm 21."

01:09:19.740 --> 01:09:21.970 align:middle line:90%
In fact, I'm 50.

01:09:21.970 --> 01:09:27.507 align:middle line:84%
But what I mean is, I
satisfy the requirement.

01:09:27.507 --> 01:09:28.590 align:middle line:90%
Do people agree with that?

01:09:28.590 --> 01:09:31.160 align:middle line:90%


01:09:31.160 --> 01:09:34.790 align:middle line:84%
There's another kind of example
like this that people give.

01:09:34.790 --> 01:09:36.620 align:middle line:84%
Imagine that the
government is going

01:09:36.620 --> 01:09:44.870 align:middle line:84%
to have a tax break for people
with two children, two or more.

01:09:44.870 --> 01:09:51.319 align:middle line:84%
And so, I claim this
tax benefit on my taxes.

01:09:51.319 --> 01:09:53.840 align:middle line:84%
And then someone from the
government comes by and says,

01:09:53.840 --> 01:09:55.220 align:middle line:90%
you claimed this benefit.

01:09:55.220 --> 01:09:56.860 align:middle line:84%
You're claiming to
have two children.

01:09:56.860 --> 01:09:58.110 align:middle line:90%
I say, I do have two children.

01:09:58.110 --> 01:09:58.943 align:middle line:90%
In fact, I have six.

01:09:58.943 --> 01:10:01.640 align:middle line:90%


01:10:01.640 --> 01:10:05.450 align:middle line:84%
This is something that people
seem to think about-- numbers,

01:10:05.450 --> 01:10:10.310 align:middle line:84%
that numbers imply
exactly the number,

01:10:10.310 --> 01:10:12.590 align:middle line:84%
but they don't entail
exactly the number.

01:10:12.590 --> 01:10:14.750 align:middle line:84%
They just imply
it very strongly.

01:10:14.750 --> 01:10:16.632 align:middle line:84%
And under the right
circumstances,

01:10:16.632 --> 01:10:18.590 align:middle line:84%
you can convince yourself
that what they really

01:10:18.590 --> 01:10:23.040 align:middle line:84%
mean, what they actually
entail is that number or more.

01:10:23.040 --> 01:10:26.160 align:middle line:84%
These kinds of circumstances
lead people to think that.

01:10:26.160 --> 01:10:26.790 align:middle line:90%
Yeah?

01:10:26.790 --> 01:10:28.980 align:middle line:84%
AUDIENCE: I wonder
if in languages

01:10:28.980 --> 01:10:31.410 align:middle line:84%
where instead of saying
things like "I am 21,"

01:10:31.410 --> 01:10:33.960 align:middle line:84%
you're saying "I have
21 years," whether it

01:10:33.960 --> 01:10:36.540 align:middle line:84%
be more likely that you
accept like oh yeah,

01:10:36.540 --> 01:10:40.960 align:middle line:84%
like if you're 50, you have
50, then you probably have 25.

01:10:40.960 --> 01:10:42.660 align:middle line:84%
Like maybe, because
we think of it

01:10:42.660 --> 01:10:44.115 align:middle line:84%
as an identity
thing [INAUDIBLE]..

01:10:44.115 --> 01:10:47.460 align:middle line:84%
NORVIN RICHARDS: It would be
interesting to study that.

01:10:47.460 --> 01:10:49.230 align:middle line:84%
So people, when I
said that I could

01:10:49.230 --> 01:10:51.570 align:middle line:84%
say to the guy at
the bar "I'm 21"--

01:10:51.570 --> 01:10:52.170 align:middle line:90%
yeah.

01:10:52.170 --> 01:10:54.360 align:middle line:84%
I'm sorry, I'm stepping
on what you're saying.

01:10:54.360 --> 01:10:54.900 align:middle line:90%
You're absolutely right.

01:10:54.900 --> 01:10:56.733 align:middle line:84%
There are languages out
there, like Spanish,

01:10:56.733 --> 01:10:58.890 align:middle line:84%
in which to say I'm 21,
you say "I have 21 years,"

01:10:58.890 --> 01:10:59.848 align:middle line:90%
or something like that.

01:10:59.848 --> 01:11:02.190 align:middle line:90%


01:11:02.190 --> 01:11:03.630 align:middle line:84%
Were people just
being polite when

01:11:03.630 --> 01:11:08.850 align:middle line:84%
I said to-- that I could say
to the guy at the bar I'm 21?

01:11:08.850 --> 01:11:10.980 align:middle line:84%
Did people want to
object that would

01:11:10.980 --> 01:11:13.230 align:middle line:90%
be a weird thing for me to say?

01:11:13.230 --> 01:11:15.730 align:middle line:84%
Some people think that would
be a weird thing for me to say.

01:11:15.730 --> 01:11:17.820 align:middle line:90%
What should I say?

01:11:17.820 --> 01:11:19.940 align:middle line:90%
I'm 50.

01:11:19.940 --> 01:11:21.410 align:middle line:90%
Yeah?

01:11:21.410 --> 01:11:23.552 align:middle line:84%
AUDIENCE: [INAUDIBLE]
like, "I'm over 21."

01:11:23.552 --> 01:11:25.010 align:middle line:84%
NORVIN RICHARDS:
Oh, "I'm over 21."

01:11:25.010 --> 01:11:27.020 align:middle line:84%
Suppose the guy has just
said to me, you have to be 21

01:11:27.020 --> 01:11:27.770 align:middle line:90%
to get inside.

01:11:27.770 --> 01:11:29.580 align:middle line:84%
First of all, notice,
he can say that.

01:11:29.580 --> 01:11:31.520 align:middle line:84%
He can say you have to
be 21 to get inside.

01:11:31.520 --> 01:11:36.440 align:middle line:84%
He doesn't mean we don't
allow 22-year-olds, right?

01:11:36.440 --> 01:11:39.320 align:middle line:84%
He means you have
to be 21 or higher.

01:11:39.320 --> 01:11:43.972 align:middle line:84%
And what I'm saying is I
satisfy your requirement.

01:11:43.972 --> 01:11:47.830 align:middle line:90%


01:11:47.830 --> 01:11:50.180 align:middle line:90%
But I take your point.

01:11:50.180 --> 01:11:52.040 align:middle line:84%
All of you go to bars
and try this out.

01:11:52.040 --> 01:11:52.540 align:middle line:90%
No--

01:11:52.540 --> 01:11:53.712 align:middle line:90%
[LAUGHS]

01:11:53.712 --> 01:11:55.420 align:middle line:84%
Oh, darn, I've been
recorded saying that.

01:11:55.420 --> 01:11:55.930 align:middle line:90%
This is bad.

01:11:55.930 --> 01:12:03.920 align:middle line:90%


01:12:03.920 --> 01:12:05.842 align:middle line:84%
Oh yeah, and "Mary ate
some of the cookies"

01:12:05.842 --> 01:12:08.300 align:middle line:84%
implies that she didn't eat
all of them, but maybe she did.

01:12:08.300 --> 01:12:10.880 align:middle line:90%


01:12:10.880 --> 01:12:14.610 align:middle line:84%
So presuppositions are
different from this.

01:12:14.610 --> 01:12:17.870 align:middle line:84%
So I can say things
like "I'm 21.

01:12:17.870 --> 01:12:18.950 align:middle line:90%
In fact, I'm 50."

01:12:18.950 --> 01:12:20.940 align:middle line:84%
Or "Mary ate some
of the cookies.

01:12:20.940 --> 01:12:24.080 align:middle line:84%
In fact, to be perfectly
frank, she ate them all,

01:12:24.080 --> 01:12:25.400 align:middle line:90%
just so you know."

01:12:25.400 --> 01:12:28.270 align:middle line:84%
And that's, perhaps, a slightly
odd way of telling you that,

01:12:28.270 --> 01:12:29.270 align:middle line:90%
but I can tell you that.

01:12:29.270 --> 01:12:31.470 align:middle line:84%
There's nothing too
surprising about it.

01:12:31.470 --> 01:12:34.010 align:middle line:84%
On the other hand, if I say
"The King of France is bald, oh,

01:12:34.010 --> 01:12:36.170 align:middle line:84%
and by the way, there
is no King of France,"

01:12:36.170 --> 01:12:38.690 align:middle line:90%
then I am very weird.

01:12:38.690 --> 01:12:39.380 align:middle line:90%
And I know.

01:12:39.380 --> 01:12:41.270 align:middle line:84%
I have spent a lifetime
researching ways

01:12:41.270 --> 01:12:42.050 align:middle line:90%
to be very weird.

01:12:42.050 --> 01:12:43.280 align:middle line:90%
This is one of the best ones.

01:12:43.280 --> 01:12:46.400 align:middle line:84%
So saying something
that presupposes--

01:12:46.400 --> 01:12:48.590 align:middle line:84%
saying A that
presupposes B, and then,

01:12:48.590 --> 01:12:51.260 align:middle line:84%
saying oh, and B
is false, that's

01:12:51.260 --> 01:12:55.100 align:middle line:84%
just like a very
peculiar thing to do.

01:12:55.100 --> 01:12:58.360 align:middle line:84%
So to summarize, we've
talked about three kinds

01:12:58.360 --> 01:13:00.230 align:middle line:90%
of relations between sentences.

01:13:00.230 --> 01:13:03.010 align:middle line:84%
So a sentence can
have entailments.

01:13:03.010 --> 01:13:05.440 align:middle line:84%
So the entailments
of a sentence A,

01:13:05.440 --> 01:13:09.100 align:middle line:90%
have to be true if A is true.

01:13:09.100 --> 01:13:10.930 align:middle line:84%
Presuppositions have
to be true for A

01:13:10.930 --> 01:13:13.340 align:middle line:90%
to be either true or false.

01:13:13.340 --> 01:13:17.410 align:middle line:84%
So "The King of France
is bald" presupposes

01:13:17.410 --> 01:13:18.830 align:middle line:90%
that there is a King of France.

01:13:18.830 --> 01:13:21.070 align:middle line:84%
So there has to be a
King of France for it

01:13:21.070 --> 01:13:25.700 align:middle line:84%
to be either true or false that
"The King of France is bald."

01:13:25.700 --> 01:13:29.087 align:middle line:84%
And then implicatures are
things that yeah, could be true.

01:13:29.087 --> 01:13:31.670 align:middle line:84%
You might be inclined to think
that they're true if A is true,

01:13:31.670 --> 01:13:32.960 align:middle line:90%
but they might be false.

01:13:32.960 --> 01:13:35.600 align:middle line:84%
And if I say A and then
say oh and by the way,

01:13:35.600 --> 01:13:40.580 align:middle line:84%
this implicature is false,
you're not distraught.

01:13:40.580 --> 01:13:42.740 align:middle line:84%
Or to put it another
way, if you're

01:13:42.740 --> 01:13:45.380 align:middle line:84%
asking about one
sentence P, is P

01:13:45.380 --> 01:13:47.060 align:middle line:84%
an entailment, or
a presupposition,

01:13:47.060 --> 01:13:48.830 align:middle line:84%
or an implicature
of A, the thing

01:13:48.830 --> 01:13:52.820 align:middle line:84%
to ask yourself is, "Does
P have to be true for A

01:13:52.820 --> 01:13:54.870 align:middle line:90%
to be either true or false?"

01:13:54.870 --> 01:13:57.710 align:middle line:84%
If so, then it's
a presupposition.

01:13:57.710 --> 01:14:00.690 align:middle line:84%
"Does P have to be true
just if A is true?"

01:14:00.690 --> 01:14:03.000 align:middle line:90%
If so, it's an entailment.

01:14:03.000 --> 01:14:05.660 align:middle line:90%
So it's not a presupposition.

01:14:05.660 --> 01:14:07.070 align:middle line:90%
It's an entailment.

01:14:07.070 --> 01:14:09.650 align:middle line:84%
And then "Does P have
to be true at all,

01:14:09.650 --> 01:14:12.650 align:middle line:84%
or is it just something you
might be inclined to think?"

01:14:12.650 --> 01:14:16.450 align:middle line:90%
So then it's an implicature.

01:14:16.450 --> 01:14:19.280 align:middle line:84%
I want to practice this
for like two minutes.

01:14:19.280 --> 01:14:23.260 align:middle line:84%
So if I say, "Bill isn't
aware that Susan is pregnant,"

01:14:23.260 --> 01:14:24.340 align:middle line:90%
that's a sentence.

01:14:24.340 --> 01:14:27.012 align:middle line:84%
And now here are two
sentences that we

01:14:27.012 --> 01:14:29.470 align:middle line:84%
can talk about the relationship
between the first sentence,

01:14:29.470 --> 01:14:31.040 align:middle line:90%
and the second, and the third.

01:14:31.040 --> 01:14:33.400 align:middle line:84%
So "Bill isn't aware
that Susan is pregnant."

01:14:33.400 --> 01:14:37.090 align:middle line:84%
What's the relationship
between that and the sentence

01:14:37.090 --> 01:14:38.320 align:middle line:90%
"Susan is pregnant?"

01:14:38.320 --> 01:14:42.518 align:middle line:84%
Is that presupposition or
entailment or implicature?

01:14:42.518 --> 01:14:43.560 align:middle line:90%
AUDIENCE: Presupposition.

01:14:43.560 --> 01:14:45.102 align:middle line:84%
NORVIN RICHARDS:
Presupposition, why?

01:14:45.102 --> 01:14:48.720 align:middle line:90%


01:14:48.720 --> 01:14:52.730 align:middle line:84%
AUDIENCE: Well because whether
Bill is aware of or not aware

01:14:52.730 --> 01:14:55.767 align:middle line:84%
that Susan is
pregnant, they both

01:14:55.767 --> 01:14:58.342 align:middle line:90%
mean Susan is pregnant versus--

01:14:58.342 --> 01:14:59.300 align:middle line:90%
NORVIN RICHARDS: Right.

01:14:59.300 --> 01:15:00.930 align:middle line:84%
That's the way to
demonstrate that.

01:15:00.930 --> 01:15:04.970 align:middle line:84%
So both the sentence--
for either "Bill

01:15:04.970 --> 01:15:06.710 align:middle line:84%
isn't aware that
Susan is pregnant"

01:15:06.710 --> 01:15:10.580 align:middle line:84%
or "Bill is aware that Susan
is pregnant" to be true,

01:15:10.580 --> 01:15:11.900 align:middle line:90%
Susan has to be pregnant.

01:15:11.900 --> 01:15:13.908 align:middle line:84%
It has to be true that
Susan is pregnant.

01:15:13.908 --> 01:15:15.950 align:middle line:84%
What about the second
sentence, or last sentence,

01:15:15.950 --> 01:15:19.370 align:middle line:84%
"You should tell Bill
that Susan is pregnant."

01:15:19.370 --> 01:15:20.570 align:middle line:90%
That's an implicature.

01:15:20.570 --> 01:15:23.420 align:middle line:84%
So there could be
circumstances under which I

01:15:23.420 --> 01:15:26.420 align:middle line:84%
would say that first sentence
on order to communicate to you

01:15:26.420 --> 01:15:30.110 align:middle line:84%
that last sentence, but it
doesn't have to be true.

01:15:30.110 --> 01:15:32.395 align:middle line:84%
It's just the kind of
thing you might conclude.

01:15:32.395 --> 01:15:33.520 align:middle line:90%
So that's a presupposition.

01:15:33.520 --> 01:15:36.100 align:middle line:90%
That's implicature.

01:15:36.100 --> 01:15:38.600 align:middle line:84%
We will do much more
of this next time.

01:15:38.600 --> 01:15:40.320 align:middle line:90%
Have a good weekend.

01:15:40.320 --> 01:15:45.000 align:middle line:90%