WEBVTT 00:00:00.000 --> 00:00:02.460 align:middle line:90% [SQUEAKING] 00:00:02.460 --> 00:00:04.428 align:middle line:90% [RUSTLING] 00:00:04.428 --> 00:00:06.396 align:middle line:90% [CLICKING] 00:00:06.396 --> 00:00:11.330 align:middle line:90% 00:00:11.330 --> 00:00:14.000 align:middle line:84% NORVIN RICHARDS: All right, why don't we start up? 00:00:14.000 --> 00:00:16.435 align:middle line:84% So we're continuing to do syntax. 00:00:16.435 --> 00:00:17.810 align:middle line:84% I think I promised you that I was 00:00:17.810 --> 00:00:20.143 align:middle line:84% going to have a look at the syllabus, and I sort of did. 00:00:20.143 --> 00:00:22.490 align:middle line:90% I think if-- 00:00:22.490 --> 00:00:25.880 align:middle line:84% I think we will probably finish syntax not today, 00:00:25.880 --> 00:00:27.830 align:middle line:90% but on Thursday. 00:00:27.830 --> 00:00:29.918 align:middle line:84% And then we will stop doing syntax. 00:00:29.918 --> 00:00:32.210 align:middle line:84% So if there's anyone who is tired of syntax, I'm sorry. 00:00:32.210 --> 00:00:33.752 align:middle line:84% Please bear with it for another week, 00:00:33.752 --> 00:00:37.550 align:middle line:84% and then there will be no more syntax for a while. 00:00:37.550 --> 00:00:39.110 align:middle line:90% This is today's "XKCD." 00:00:39.110 --> 00:00:41.882 align:middle line:84% It's not true, as far as I know, or maybe I'm 00:00:41.882 --> 00:00:43.340 align:middle line:84% the only one who didn't get a word. 00:00:43.340 --> 00:00:45.173 align:middle line:84% That's-- I guess that's another possibility. 00:00:45.173 --> 00:00:50.960 align:middle line:84% We'll have to ask-- ask your TAs in the sessions tomorrow. 00:00:50.960 --> 00:00:54.038 align:middle line:90% 00:00:54.038 --> 00:01:02.560 align:middle line:84% OK, so one topic that we keep coming back to is the fact-- 00:01:02.560 --> 00:01:03.600 align:middle line:90% seems to be a fact-- 00:01:03.600 --> 00:01:05.940 align:middle line:84% that although there are many kinds 00:01:05.940 --> 00:01:08.340 align:middle line:84% of languages in the world, there are not as many 00:01:08.340 --> 00:01:09.580 align:middle line:90% as you could imagine. 00:01:09.580 --> 00:01:12.060 align:middle line:84% So various types of languages, places 00:01:12.060 --> 00:01:16.080 align:middle line:84% where we've seen that languages make choices with respect 00:01:16.080 --> 00:01:19.170 align:middle line:84% to how they do this or that or the other syntactic thing. 00:01:19.170 --> 00:01:22.440 align:middle line:84% But you don't find all of the choices that are logically 00:01:22.440 --> 00:01:24.040 align:middle line:90% possible. 00:01:24.040 --> 00:01:26.880 align:middle line:84% So for example, we've seen that there are languages 00:01:26.880 --> 00:01:29.310 align:middle line:84% that move their wh-phrases to the left periphery 00:01:29.310 --> 00:01:31.740 align:middle line:84% of the clause and there are languages that just leave 00:01:31.740 --> 00:01:33.697 align:middle line:84% their wh-phrases where they are, but there 00:01:33.697 --> 00:01:35.280 align:middle line:84% don't appear to be languages that move 00:01:35.280 --> 00:01:36.867 align:middle line:90% their wh-phrases anywhere else. 00:01:36.867 --> 00:01:39.450 align:middle line:84% There aren't languages that move them to the end of the clause 00:01:39.450 --> 00:01:41.533 align:middle line:84% or that put them in the exact middle of the clause 00:01:41.533 --> 00:01:42.520 align:middle line:90% or anything like that. 00:01:42.520 --> 00:01:45.120 align:middle line:84% So there's variation, but not as much variation 00:01:45.120 --> 00:01:47.670 align:middle line:90% as you could imagine. 00:01:47.670 --> 00:01:52.110 align:middle line:84% I want to talk about another way, another case of that kind, 00:01:52.110 --> 00:01:56.790 align:middle line:84% a place where there's more than one way to build a language, 00:01:56.790 --> 00:02:01.120 align:middle line:84% but there aren't as many ways as there could be. 00:02:01.120 --> 00:02:03.900 align:middle line:84% So we're going to talk about the embedded clause 00:02:03.900 --> 00:02:06.330 align:middle line:84% of this sentence, "I thought that Mary ate sushi 00:02:06.330 --> 00:02:07.170 align:middle line:90% with chopsticks." 00:02:07.170 --> 00:02:12.090 align:middle line:84% And actually, let's-- just as a class exercise, 00:02:12.090 --> 00:02:14.910 align:middle line:84% let's diagram that embedded clause. 00:02:14.910 --> 00:02:16.710 align:middle line:84% So don't worry about the matrix clause. 00:02:16.710 --> 00:02:18.840 align:middle line:90% So that-- what's her name? 00:02:18.840 --> 00:02:23.710 align:middle line:84% "Mary--" oh dear, I'm going to make it future-- 00:02:23.710 --> 00:02:31.450 align:middle line:84% "will eat sushi with chopsticks." 00:02:31.450 --> 00:02:35.020 align:middle line:90% So let's draw a tree for this. 00:02:35.020 --> 00:02:38.230 align:middle line:84% Maybe we could start by labeling everything. 00:02:38.230 --> 00:02:39.250 align:middle line:90% What's "chopsticks?" 00:02:39.250 --> 00:02:40.905 align:middle line:90% What type of word is that? 00:02:40.905 --> 00:02:41.530 align:middle line:90% AUDIENCE: Noun. 00:02:41.530 --> 00:02:42.738 align:middle line:90% NORVIN RICHARDS: It's a noun. 00:02:42.738 --> 00:02:43.883 align:middle line:90% And "with?" 00:02:43.883 --> 00:02:44.800 align:middle line:90% AUDIENCE: Preposition. 00:02:44.800 --> 00:02:45.340 align:middle line:90% NORVIN RICHARDS: Preposition. 00:02:45.340 --> 00:02:46.480 align:middle line:90% And "sushi?" 00:02:46.480 --> 00:02:48.370 align:middle line:84% Noun-- a particularly delicious noun. 00:02:48.370 --> 00:02:50.290 align:middle line:90% And "eat" is a verb. 00:02:50.290 --> 00:02:54.365 align:middle line:84% And "will" is a tense, and "Mary" is a noun-- one 00:02:54.365 --> 00:02:55.240 align:middle line:90% of my favorite nouns. 00:02:55.240 --> 00:02:56.650 align:middle line:90% And "that?" 00:02:56.650 --> 00:02:58.697 align:middle line:84% It's a complementizer, C, excellent. 00:02:58.697 --> 00:03:01.030 align:middle line:84% And if anybody is sitting there quietly thinking, "What? 00:03:01.030 --> 00:03:02.950 align:middle line:90% Where did all that come from?" 00:03:02.950 --> 00:03:04.900 align:middle line:84% ask your TAs, or we can talk about it now. 00:03:04.900 --> 00:03:07.510 align:middle line:84% Is there anything on here that people are like, wait, no. 00:03:07.510 --> 00:03:10.130 align:middle line:90% Why did you do that? 00:03:10.130 --> 00:03:10.630 align:middle line:90% No? 00:03:10.630 --> 00:03:11.130 align:middle line:90% OK. 00:03:11.130 --> 00:03:13.135 align:middle line:84% All right, so now let's begin merging things. 00:03:13.135 --> 00:03:15.670 align:middle line:90% 00:03:15.670 --> 00:03:18.340 align:middle line:84% Somebody give me two nodes on this tree 00:03:18.340 --> 00:03:20.444 align:middle line:90% that I ought to merge. 00:03:20.444 --> 00:03:20.944 align:middle line:90% Faith? 00:03:20.944 --> 00:03:22.260 align:middle line:84% AUDIENCE: "With" and "chopsticks" 00:03:22.260 --> 00:03:23.968 align:middle line:84% NORVIN RICHARDS: "With" and "chopsticks." 00:03:23.968 --> 00:03:25.940 align:middle line:84% And what label should the resulting thing have? 00:03:25.940 --> 00:03:26.450 align:middle line:90% AUDIENCE: Preposition. 00:03:26.450 --> 00:03:28.220 align:middle line:84% NORVIN RICHARDS: Yeah, this is a prepositional phrase. 00:03:28.220 --> 00:03:30.470 align:middle line:84% And because this is the highest thing with the node N, 00:03:30.470 --> 00:03:34.430 align:middle line:84% I'll give it a P too, so that's now a noun phrase yeah? 00:03:34.430 --> 00:03:35.093 align:middle line:90% Joseph? 00:03:35.093 --> 00:03:35.760 align:middle line:90% AUDIENCE: "Eat." 00:03:35.760 --> 00:03:37.280 align:middle line:84% NORVIN RICHARDS: "Eat" and "sushi." 00:03:37.280 --> 00:03:39.115 align:middle line:84% And what label should I give this? 00:03:39.115 --> 00:03:39.740 align:middle line:90% AUDIENCE: Verb? 00:03:39.740 --> 00:03:41.060 align:middle line:84% NORVIN RICHARDS: Yeah, it's going to be a V bar in the end. 00:03:41.060 --> 00:03:42.390 align:middle line:90% I'll just call it a V for now. 00:03:42.390 --> 00:03:46.320 align:middle line:84% And now this is an NP because its label didn't project. 00:03:46.320 --> 00:03:46.820 align:middle line:90% Yeah? 00:03:46.820 --> 00:03:48.080 align:middle line:90% What else should I merge? 00:03:48.080 --> 00:03:49.232 align:middle line:90% AUDIENCE: "That Mary?" 00:03:49.232 --> 00:03:50.440 align:middle line:90% NORVIN RICHARDS: "That Mary?" 00:03:50.440 --> 00:03:52.340 align:middle line:90% Well, I could, yeah. 00:03:52.340 --> 00:03:55.434 align:middle line:90% And what would I project there? 00:03:55.434 --> 00:03:57.103 align:middle line:84% AUDIENCE: A complementizer phrase. 00:03:57.103 --> 00:03:59.645 align:middle line:84% NORVIN RICHARDS: This is going to be a complementizer phrase? 00:03:59.645 --> 00:04:00.350 align:middle line:90% OK, possibly. 00:04:00.350 --> 00:04:02.000 align:middle line:84% And then this would be a noun phrase. 00:04:02.000 --> 00:04:04.170 align:middle line:90% We did that, yeah. 00:04:04.170 --> 00:04:06.480 align:middle line:84% Any other nodes here that I ought to merge? 00:04:06.480 --> 00:04:09.480 align:middle line:84% Including nodes that we've made in the course of doing this. 00:04:09.480 --> 00:04:10.140 align:middle line:90% Yes? 00:04:10.140 --> 00:04:11.835 align:middle line:84% AUDIENCE: "Eats sushi with chopsticks." 00:04:11.835 --> 00:04:14.252 align:middle line:84% NORVIN RICHARDS: Yeah, "eats sushi" and "with chopsticks." 00:04:14.252 --> 00:04:16.079 align:middle line:90% And what label will that have? 00:04:16.079 --> 00:04:17.470 align:middle line:90% That's a verb phrase. 00:04:17.470 --> 00:04:17.970 align:middle line:90% Cool. 00:04:17.970 --> 00:04:19.095 align:middle line:90% And so now this is a V bar. 00:04:19.095 --> 00:04:21.380 align:middle line:90% Joseph was right. 00:04:21.380 --> 00:04:24.420 align:middle line:90% Anything else I should merge? 00:04:24.420 --> 00:04:25.020 align:middle line:90% Yeah, sorry. 00:04:25.020 --> 00:04:26.596 align:middle line:90% Go ahead, Faith." 00:04:26.596 --> 00:04:28.490 align:middle line:84% AUDIENCE: The two words that should be-- 00:04:28.490 --> 00:04:28.990 align:middle line:84% NORVIN RICHARDS: I think we should 00:04:28.990 --> 00:04:30.100 align:middle line:90% merge these two things, yes. 00:04:30.100 --> 00:04:32.730 align:middle line:84% And we should give, as you just said, the whole thing the label 00:04:32.730 --> 00:04:33.772 align:middle line:90% T. Yeah, you're right. 00:04:33.772 --> 00:04:35.230 align:middle line:84% Is that what you were going to say? 00:04:35.230 --> 00:04:36.868 align:middle line:90% Yeah. 00:04:36.868 --> 00:04:37.910 align:middle line:90% What else should I merge? 00:04:37.910 --> 00:04:38.450 align:middle line:90% Yes? 00:04:38.450 --> 00:04:39.230 align:middle line:90% AUDIENCE: Just a question. 00:04:39.230 --> 00:04:40.180 align:middle line:90% NORVIN RICHARDS: Yeah. 00:04:40.180 --> 00:04:41.638 align:middle line:84% AUDIENCE: Could you quickly explain 00:04:41.638 --> 00:04:43.940 align:middle line:90% the bars again, versus the-- 00:04:43.940 --> 00:04:46.433 align:middle line:84% NORVIN RICHARDS: Oh, so here we have three nodes. 00:04:46.433 --> 00:04:47.600 align:middle line:90% This is a nice example here. 00:04:47.600 --> 00:04:49.225 align:middle line:84% We have three nodes, all with the label 00:04:49.225 --> 00:04:53.900 align:middle line:84% V. P is just the name for the highest one, yeah. 00:04:53.900 --> 00:04:56.510 align:middle line:84% And the lowest one doesn't get a mark, 00:04:56.510 --> 00:04:58.310 align:middle line:84% or sometimes you'll see people put 00:04:58.310 --> 00:05:01.910 align:middle line:84% a raised zero just to mark the fact that it's the lowest form. 00:05:01.910 --> 00:05:04.380 align:middle line:84% I haven't been bothering with that. 00:05:04.380 --> 00:05:07.400 align:middle line:84% And then everything else is given the label bar. 00:05:07.400 --> 00:05:08.180 align:middle line:90% It's-- yeah. 00:05:08.180 --> 00:05:09.340 align:middle line:90% Yeah, yeah. 00:05:09.340 --> 00:05:13.760 align:middle line:84% Just higher than zero and lower than P. Syntactic math-- 00:05:13.760 --> 00:05:18.230 align:middle line:84% we count 0, bar, P. That's how we count. 00:05:18.230 --> 00:05:20.510 align:middle line:84% What other things should I merge with each other? 00:05:20.510 --> 00:05:29.130 align:middle line:90% 00:05:29.130 --> 00:05:30.832 align:middle line:90% Does anybody remember the EPP? 00:05:30.832 --> 00:05:34.690 align:middle line:90% 00:05:34.690 --> 00:05:37.870 align:middle line:90% One of my favorite Ps. 00:05:37.870 --> 00:05:42.507 align:middle line:84% It says that TP must have a specifier. 00:05:42.507 --> 00:05:46.980 align:middle line:90% 00:05:46.980 --> 00:05:52.650 align:middle line:84% So it's responsible for the fact that you can't just say-- 00:05:52.650 --> 00:06:04.300 align:middle line:84% you can say things like "is obvious that syntax is fun." 00:06:04.300 --> 00:06:07.480 align:middle line:84% You put this "it" here and this "it" 00:06:07.480 --> 00:06:09.040 align:middle line:84% is what we were calling an expletive. 00:06:09.040 --> 00:06:12.850 align:middle line:84% It's a meaningless thing that you put there so 00:06:12.850 --> 00:06:14.560 align:middle line:90% that TP can have a specifier. 00:06:14.560 --> 00:06:16.810 align:middle line:84% If I were going to draw a tree for "it is obvious 00:06:16.810 --> 00:06:21.780 align:middle line:84% that syntax is fun," it would have a TP 00:06:21.780 --> 00:06:26.130 align:middle line:84% and "it" would be in the specifier of that. 00:06:26.130 --> 00:06:28.690 align:middle line:84% We'd have a verb phrase, "is obvious," blah de blah. 00:06:28.690 --> 00:06:31.210 align:middle line:90% 00:06:31.210 --> 00:06:33.440 align:middle line:84% So I'd have a tree sort of like that one. 00:06:33.440 --> 00:06:33.940 align:middle line:90% Joseph? 00:06:33.940 --> 00:06:35.580 align:middle line:84% AUDIENCE: What does EPP stand for? 00:06:35.580 --> 00:06:37.440 align:middle line:84% NORVIN RICHARDS: Well, I encouraged you not 00:06:37.440 --> 00:06:38.970 align:middle line:84% to worry about the answer to that, 00:06:38.970 --> 00:06:42.390 align:middle line:84% but the answer is that it stands for "extended projection 00:06:42.390 --> 00:06:44.890 align:middle line:90% principle." 00:06:44.890 --> 00:06:46.980 align:middle line:84% And it is also a parameter-- sorry, 00:06:46.980 --> 00:06:49.560 align:middle line:84% let me see if I can spell "principle" correctly. 00:06:49.560 --> 00:06:51.535 align:middle line:84% It is also a parameter in that English has it-- 00:06:51.535 --> 00:06:53.910 align:middle line:84% and there are other languages that have it, like French-- 00:06:53.910 --> 00:06:55.493 align:middle line:84% but it's actually not all that common. 00:06:55.493 --> 00:06:59.860 align:middle line:84% There are plenty of languages out there that don't have this. 00:06:59.860 --> 00:07:03.870 align:middle line:84% So there's a general principle that TP must have a specifier. 00:07:03.870 --> 00:07:06.900 align:middle line:84% That is, there must be a structure like this one 00:07:06.900 --> 00:07:10.590 align:middle line:84% where there's a TP that has, as its daughters, something-- 00:07:10.590 --> 00:07:12.090 align:middle line:90% it's usually a noun phrase-- 00:07:12.090 --> 00:07:14.600 align:middle line:90% and then a T bar. 00:07:14.600 --> 00:07:17.960 align:middle line:90% So we're not yet there. 00:07:17.960 --> 00:07:20.900 align:middle line:84% This TP doesn't have a specifier, so it needs one. 00:07:20.900 --> 00:07:25.150 align:middle line:90% 00:07:25.150 --> 00:07:25.750 align:middle line:90% Yes? 00:07:25.750 --> 00:07:29.500 align:middle line:84% AUDIENCE: Can "that" be a specifier? 00:07:29.500 --> 00:07:31.540 align:middle line:84% NORVIN RICHARDS: So in a sentence 00:07:31.540 --> 00:07:37.310 align:middle line:84% where a CP is the subject of a clause-- 00:07:37.310 --> 00:07:41.560 align:middle line:84% so in a sentence like "that syntax 00:07:41.560 --> 00:07:49.720 align:middle line:84% is fun is obvious," where the subject of the predicate 00:07:49.720 --> 00:07:52.690 align:middle line:84% "is obvious" is this CP, "that syntax 00:07:52.690 --> 00:07:57.100 align:middle line:84% is fun," where "that" is its head, then yeah. 00:07:57.100 --> 00:08:00.610 align:middle line:84% So if this were a clause, then yeah, 00:08:00.610 --> 00:08:02.860 align:middle line:90% it could be the specifier of TP. 00:08:02.860 --> 00:08:03.475 align:middle line:90% Joseph? 00:08:03.475 --> 00:08:05.920 align:middle line:84% AUDIENCE: Could "Mary" not merge with the T bar, 00:08:05.920 --> 00:08:08.535 align:middle line:90% and then T merges with-- 00:08:08.535 --> 00:08:09.910 align:middle line:84% NORVIN RICHARDS: I think we might 00:08:09.910 --> 00:08:11.493 align:middle line:84% want to think about it that way, yeah. 00:08:11.493 --> 00:08:13.990 align:middle line:84% We might want to make "Mary" the specifier of TP, 00:08:13.990 --> 00:08:15.970 align:middle line:90% satisfying the EPP, yeah. 00:08:15.970 --> 00:08:19.750 align:middle line:84% And then all we've got left is a C, 00:08:19.750 --> 00:08:26.500 align:middle line:84% and we can make that C the sister of this TP, yeah. 00:08:26.500 --> 00:08:31.173 align:middle line:84% Is that a tree that people are not too unnerved by? 00:08:31.173 --> 00:08:32.590 align:middle line:84% All of you have all the nerves you 00:08:32.590 --> 00:08:35.510 align:middle line:90% had before you looked at it. 00:08:35.510 --> 00:08:36.240 align:middle line:90% Yeah? 00:08:36.240 --> 00:08:36.740 align:middle line:90% OK. 00:08:36.740 --> 00:08:42.217 align:middle line:84% All right, so there is a tree for the embedded clause, 00:08:42.217 --> 00:08:44.300 align:middle line:84% the boldfaced embedded clause "that Mary ate sushi 00:08:44.300 --> 00:08:45.350 align:middle line:90% with chopsticks." 00:08:45.350 --> 00:08:48.290 align:middle line:84% And here is hopefully the same tree. 00:08:48.290 --> 00:08:49.640 align:middle line:90% Oh, I put it in a "could," yeah. 00:08:49.640 --> 00:08:54.143 align:middle line:84% So "that Mary could eat sushi with chopsticks," same deal. 00:08:54.143 --> 00:08:55.590 align:middle line:90% OK? 00:08:55.590 --> 00:08:56.760 align:middle line:90% All right. 00:08:56.760 --> 00:08:58.920 align:middle line:84% Now let me call your attention to a fact 00:08:58.920 --> 00:09:02.160 align:middle line:84% about this tree, which some of you may already have noticed. 00:09:02.160 --> 00:09:05.160 align:middle line:84% A general rule when you're drawing a tree for English, 00:09:05.160 --> 00:09:07.590 align:middle line:84% that if you have a head and the head has a sister, 00:09:07.590 --> 00:09:10.320 align:middle line:90% the head goes before the sister. 00:09:10.320 --> 00:09:13.620 align:middle line:84% So we have prepositional phrases like "with chopsticks" 00:09:13.620 --> 00:09:17.220 align:middle line:84% and verb phrases that have-- if they have an object, 00:09:17.220 --> 00:09:19.590 align:middle line:84% you get the verb before the object, "eat sushi." 00:09:19.590 --> 00:09:22.500 align:middle line:84% And T, like "could" or "will," which we have on the board, 00:09:22.500 --> 00:09:24.510 align:middle line:84% precedes its complement, the verb phrase. 00:09:24.510 --> 00:09:27.000 align:middle line:84% And C, the "that" which is up there, 00:09:27.000 --> 00:09:28.810 align:middle line:84% precedes its complement, which is the TP. 00:09:28.810 --> 00:09:30.060 align:middle line:90% We have all these blue arrows. 00:09:30.060 --> 00:09:31.260 align:middle line:84% On the tree, the blue arrows are just 00:09:31.260 --> 00:09:32.760 align:middle line:84% meant to-- you don't have to draw them if you're 00:09:32.760 --> 00:09:33.760 align:middle line:90% drawing an English tree. 00:09:33.760 --> 00:09:36.563 align:middle line:84% They're just there to dramatically represent the fact 00:09:36.563 --> 00:09:37.980 align:middle line:84% that here we have all these heads, 00:09:37.980 --> 00:09:39.750 align:middle line:84% and they're preceding their complements. 00:09:39.750 --> 00:09:41.628 align:middle line:90% Yeah. 00:09:41.628 --> 00:09:43.170 align:middle line:84% What would English look like if heads 00:09:43.170 --> 00:09:45.030 align:middle line:90% followed their complements? 00:09:45.030 --> 00:09:48.400 align:middle line:84% Well, you'd get weird orders like that one-- 00:09:48.400 --> 00:09:51.230 align:middle line:84% "Mary chopsticks with sushi eat could that," 00:09:51.230 --> 00:09:52.105 align:middle line:90% which is not English. 00:09:52.105 --> 00:09:54.910 align:middle line:90% 00:09:54.910 --> 00:09:57.190 align:middle line:90% But it is Japanese. 00:09:57.190 --> 00:10:01.810 align:middle line:84% So in Japanese, the way you say "that Mary could eat sushi 00:10:01.810 --> 00:10:03.940 align:middle line:84% with chopsticks" is literally something 00:10:03.940 --> 00:10:07.930 align:middle line:84% like "Mary chopsticks with sushi eat good that." 00:10:07.930 --> 00:10:10.000 align:middle line:84% If any of you have studied Japanese 00:10:10.000 --> 00:10:11.830 align:middle line:84% or are thinking about studying Japanese, 00:10:11.830 --> 00:10:13.210 align:middle line:84% be aware that this is something you'll 00:10:13.210 --> 00:10:15.252 align:middle line:84% have to learn to cope with, saying your sentences 00:10:15.252 --> 00:10:17.110 align:middle line:90% in this different order. 00:10:17.110 --> 00:10:19.150 align:middle line:90% So you say, [SPEAKING JAPANESE]. 00:10:19.150 --> 00:10:22.180 align:middle line:90% 00:10:22.180 --> 00:10:24.550 align:middle line:84% And not just Japanese, but well, lots 00:10:24.550 --> 00:10:26.000 align:middle line:90% of other languages as well. 00:10:26.000 --> 00:10:29.200 align:middle line:84% So that's the basic word order for Tibetan and for Korean 00:10:29.200 --> 00:10:31.103 align:middle line:84% and for Navajo and Basque and Chaha 00:10:31.103 --> 00:10:32.770 align:middle line:84% and a zillion other languages out there. 00:10:32.770 --> 00:10:34.978 align:middle line:84% It's, cross linguistically, a very common word order. 00:10:34.978 --> 00:10:37.055 align:middle line:84% In fact, if you just count languages, 00:10:37.055 --> 00:10:38.680 align:middle line:84% it might be the most common word order. 00:10:38.680 --> 00:10:44.280 align:middle line:84% It's slightly more common than the English style word order. 00:10:44.280 --> 00:10:50.070 align:middle line:84% OK, so here's a single switch that you can flick, right? 00:10:50.070 --> 00:10:51.960 align:middle line:84% Do your heads precede your complements 00:10:51.960 --> 00:10:53.550 align:middle line:84% or do they follow your complements? 00:10:53.550 --> 00:10:56.130 align:middle line:84% So in English, the heads precede the complements. 00:10:56.130 --> 00:10:58.170 align:middle line:84% In Japanese and lots of other languages, 00:10:58.170 --> 00:11:00.262 align:middle line:90% they follow the complements. 00:11:00.262 --> 00:11:02.220 align:middle line:84% There's one basic difference between languages. 00:11:02.220 --> 00:11:04.880 align:middle line:90% 00:11:04.880 --> 00:11:12.170 align:middle line:84% Now it used to be that this was the point in the class where 00:11:12.170 --> 00:11:14.357 align:middle line:84% I would tell you that, and I would wait just 00:11:14.357 --> 00:11:16.190 align:middle line:84% long enough for you to be impressed by that, 00:11:16.190 --> 00:11:18.080 align:middle line:84% and then I would quickly change the subject 00:11:18.080 --> 00:11:21.170 align:middle line:84% and hope that none of you spoke German. 00:11:21.170 --> 00:11:23.180 align:middle line:84% Because there are languages out there, 00:11:23.180 --> 00:11:25.650 align:middle line:84% sadly, in which some heads are initial and other heads 00:11:25.650 --> 00:11:26.150 align:middle line:90% are final. 00:11:26.150 --> 00:11:28.010 align:middle line:84% That is, some heads precede their complements and others 00:11:28.010 --> 00:11:29.420 align:middle line:90% follow their complements. 00:11:29.420 --> 00:11:31.770 align:middle line:90% German is such a language. 00:11:31.770 --> 00:11:36.080 align:middle line:84% So for example, languages with mixed headedness-- in German, 00:11:36.080 --> 00:11:39.830 align:middle line:84% here's German for "that Mary could eat sushi 00:11:39.830 --> 00:11:40.550 align:middle line:90% with chopsticks." 00:11:40.550 --> 00:11:43.660 align:middle line:84% And you can see the German complementizer precedes 00:11:43.660 --> 00:11:44.160 align:middle line:90% the clause. 00:11:44.160 --> 00:11:46.220 align:middle line:84% So the German compromiser for that is "dass," 00:11:46.220 --> 00:11:49.430 align:middle line:84% and it goes before the clause, just like in English. 00:11:49.430 --> 00:11:52.930 align:middle line:84% And German has prepositions, so "with chopsticks" 00:11:52.930 --> 00:11:55.370 align:middle line:90% is "mit Stäbchen." 00:11:55.370 --> 00:11:56.920 align:middle line:84% So "with" goes before "chopsticks." 00:11:56.920 --> 00:11:58.670 align:middle line:84% You know, German has prepositional phrases 00:11:58.670 --> 00:12:00.140 align:middle line:90% just like English. 00:12:00.140 --> 00:12:03.710 align:middle line:84% But German verbs and German tense, whatever 00:12:03.710 --> 00:12:05.840 align:middle line:84% we put in T, these kinds of auxiliaries, 00:12:05.840 --> 00:12:07.340 align:middle line:84% at least in this kind of clause-- 00:12:07.340 --> 00:12:08.630 align:middle line:90% we'll come back to this-- 00:12:08.630 --> 00:12:10.580 align:middle line:84% they come after their complements. 00:12:10.580 --> 00:12:13.465 align:middle line:84% So if you're saying in German, "I 00:12:13.465 --> 00:12:15.590 align:middle line:84% thought that Mary could eat sushi with chopsticks," 00:12:15.590 --> 00:12:16.970 align:middle line:84% the word order is literally going 00:12:16.970 --> 00:12:21.170 align:middle line:84% to be something like "I thought that Mary with chopsticks sushi 00:12:21.170 --> 00:12:23.270 align:middle line:90% eat could." 00:12:23.270 --> 00:12:27.680 align:middle line:84% Mark Twain has a great essay called "The Awful German 00:12:27.680 --> 00:12:34.280 align:middle line:84% Language" in which he says many hilariously partially accurate 00:12:34.280 --> 00:12:38.480 align:middle line:84% things about German, one of them being that-- he says something 00:12:38.480 --> 00:12:41.870 align:middle line:84% like, when a German dives into a sentence, that is the last you 00:12:41.870 --> 00:12:45.170 align:middle line:84% will see of him until he emerges from the other side 00:12:45.170 --> 00:12:47.900 align:middle line:84% of the Atlantic with his verb in his mouth. 00:12:47.900 --> 00:12:49.580 align:middle line:84% Because he's making fun of the fact 00:12:49.580 --> 00:12:52.340 align:middle line:84% that German often has the verb at the end of the sentence, 00:12:52.340 --> 00:12:55.640 align:middle line:84% and goes on and on about how it's 00:12:55.640 --> 00:12:58.590 align:middle line:84% very easy to forget what exactly is being done to these things. 00:12:58.590 --> 00:13:00.440 align:middle line:84% They're just being named, and you find out 00:13:00.440 --> 00:13:03.210 align:middle line:90% the verb weeks later. 00:13:03.210 --> 00:13:03.790 align:middle line:90% So, yeah. 00:13:03.790 --> 00:13:06.290 align:middle line:84% So in English, heads precede their complements; in Japanese, 00:13:06.290 --> 00:13:09.190 align:middle line:84% heads follow their complements And in other languages, 00:13:09.190 --> 00:13:11.915 align:middle line:84% some other languages like German, you get both. 00:13:11.915 --> 00:13:13.790 align:middle line:84% There are some heads that precede complements 00:13:13.790 --> 00:13:16.850 align:middle line:84% and others that follow their complements, OK? 00:13:16.850 --> 00:13:17.620 align:middle line:90% Yeah, Faith. 00:13:17.620 --> 00:13:19.795 align:middle line:84% AUDIENCE: Is there any rule for which heads precede 00:13:19.795 --> 00:13:20.530 align:middle line:90% and which follow? 00:13:20.530 --> 00:13:21.070 align:middle line:90% NORVIN RICHARDS: Ah, good. 00:13:21.070 --> 00:13:21.940 align:middle line:90% Good question. 00:13:21.940 --> 00:13:23.560 align:middle line:84% That was the next question I was going 00:13:23.560 --> 00:13:24.820 align:middle line:90% to ask myself, rhetorically. 00:13:24.820 --> 00:13:27.580 align:middle line:84% Thank you for asking me non-rhetorically. 00:13:27.580 --> 00:13:31.420 align:middle line:84% You might wonder, OK, fine, so do we just 00:13:31.420 --> 00:13:33.370 align:middle line:84% have to say, for every head in every language, 00:13:33.370 --> 00:13:35.140 align:middle line:84% this head precedes, this head follows? 00:13:35.140 --> 00:13:36.130 align:middle line:90% Is there any rule? 00:13:36.130 --> 00:13:38.500 align:middle line:84% Do we get to say anything more interesting than that? 00:13:38.500 --> 00:13:40.270 align:middle line:84% It turns out that there are kinds 00:13:40.270 --> 00:13:46.310 align:middle line:84% of systems that don't exist, which is kind of interesting. 00:13:46.310 --> 00:13:51.820 align:middle line:84% So let's concentrate on the heads T and V. 00:13:51.820 --> 00:13:54.370 align:middle line:84% So we're going to look at embedded clauses 00:13:54.370 --> 00:13:56.830 align:middle line:84% like this one, where you've got something in tense, 00:13:56.830 --> 00:13:59.080 align:middle line:84% an auxiliary of some kind, and you've got a verb, 00:13:59.080 --> 00:14:00.910 align:middle line:90% and there's also an object. 00:14:00.910 --> 00:14:04.450 align:middle line:84% And we're going to look at the ordering of those two heads, 00:14:04.450 --> 00:14:06.430 align:middle line:84% T with respect to the verb phrase, 00:14:06.430 --> 00:14:09.340 align:middle line:84% and the verb with respect to the object. 00:14:09.340 --> 00:14:12.330 align:middle line:84% If you do that, here's what you find. 00:14:12.330 --> 00:14:15.930 align:middle line:84% You get languages, like English, in which "has" 00:14:15.930 --> 00:14:18.720 align:middle line:84% and "read" both precede their complements. 00:14:18.720 --> 00:14:22.840 align:middle line:84% You get languages like German, in which "read" and "has" 00:14:22.840 --> 00:14:24.090 align:middle line:90% both follow their complements. 00:14:24.090 --> 00:14:26.760 align:middle line:90% That's why they're both blue. 00:14:26.760 --> 00:14:29.640 align:middle line:84% You get languages like West Flemish, 00:14:29.640 --> 00:14:33.240 align:middle line:84% which is a language closely related to Dutch-- 00:14:33.240 --> 00:14:35.482 align:middle line:90% spoken, I assume, in West Flem-- 00:14:35.482 --> 00:14:36.720 align:middle line:90% [STUDENT SNEEZING] 00:14:36.720 --> 00:14:39.180 align:middle line:84% --in which the auxiliary-- bless you-- 00:14:39.180 --> 00:14:42.150 align:middle line:84% the auxiliary precedes the verb phrase, and in which 00:14:42.150 --> 00:14:43.990 align:middle line:90% the verb follows the object. 00:14:43.990 --> 00:14:47.460 align:middle line:84% So in West Flemish, you say that "John wants a house to buy." 00:14:47.460 --> 00:14:50.700 align:middle line:90% 00:14:50.700 --> 00:14:51.690 align:middle line:90% I wasn't able to do-- 00:14:51.690 --> 00:14:53.160 align:middle line:84% I don't speak West Flemish, so I wasn't 00:14:53.160 --> 00:14:54.780 align:middle line:84% able to do "that John has read a book." 00:14:54.780 --> 00:14:56.697 align:middle line:84% I guess I should make all of these "John wants 00:14:56.697 --> 00:14:58.180 align:middle line:90% to buy a house." 00:14:58.180 --> 00:15:00.900 align:middle line:84% So OK, so you get languages like English, in which both of those 00:15:00.900 --> 00:15:03.997 align:middle line:84% precede their complements, both the T and the V 00:15:03.997 --> 00:15:05.080 align:middle line:90% precede their complements. 00:15:05.080 --> 00:15:07.230 align:middle line:84% You get languages like German, in which they both follow. 00:15:07.230 --> 00:15:08.730 align:middle line:84% You get languages like West Flemish, 00:15:08.730 --> 00:15:13.390 align:middle line:84% in which the auxiliary precedes and the verb follows. 00:15:13.390 --> 00:15:17.580 align:middle line:84% But you never get the fourth imaginable kind of language-- 00:15:17.580 --> 00:15:18.510 align:middle line:90% not ever. 00:15:18.510 --> 00:15:20.520 align:middle line:90% People have looked quite hard. 00:15:20.520 --> 00:15:22.548 align:middle line:84% Where I, again, am using silly diacritics 00:15:22.548 --> 00:15:24.840 align:middle line:84% to emphasize the fact that this language doesn't exist. 00:15:24.840 --> 00:15:26.070 align:middle line:90% I'm making it up, yeah. 00:15:26.070 --> 00:15:27.930 align:middle line:84% So there are no languages in which you say 00:15:27.930 --> 00:15:29.640 align:middle line:90% "that John read the book has." 00:15:29.640 --> 00:15:34.300 align:middle line:84% That's not a possible human language, for some reason. 00:15:34.300 --> 00:15:35.968 align:middle line:84% So we get English, where those heads 00:15:35.968 --> 00:15:37.260 align:middle line:90% both precede their complements. 00:15:37.260 --> 00:15:39.593 align:middle line:84% We get German, where they both follow their complements. 00:15:39.593 --> 00:15:43.270 align:middle line:84% We get West Flemish, where the lower head follows 00:15:43.270 --> 00:15:44.520 align:middle line:90% the complement, is head final. 00:15:44.520 --> 00:15:47.820 align:middle line:84% So the verb follows the object, but the higher head, T, 00:15:47.820 --> 00:15:49.020 align:middle line:90% precedes its complement. 00:15:49.020 --> 00:15:50.025 align:middle line:90% The T is head initial. 00:15:50.025 --> 00:15:52.470 align:middle line:84% You never get the mirror image of West Flemish. 00:15:52.470 --> 00:15:55.470 align:middle line:90% That doesn't happen. 00:15:55.470 --> 00:15:58.920 align:middle line:84% It not only doesn't happen, but it sometimes 00:15:58.920 --> 00:16:01.870 align:middle line:84% fails to happen in kind of interesting ways. 00:16:01.870 --> 00:16:04.110 align:middle line:90% Here's a fact about Finnish-- 00:16:04.110 --> 00:16:09.690 align:middle line:84% Finnish word order is mostly sort of English-like. 00:16:09.690 --> 00:16:12.030 align:middle line:84% T and V both precede their complements. 00:16:12.030 --> 00:16:15.600 align:middle line:84% But just if you're asking wh-questions, for some reason, 00:16:15.600 --> 00:16:18.220 align:middle line:84% Finnish word order becomes quite random. 00:16:18.220 --> 00:16:22.050 align:middle line:84% So you can ask questions like "When would Jussi 00:16:22.050 --> 00:16:23.040 align:middle line:90% have written a novel?" 00:16:23.040 --> 00:16:25.650 align:middle line:84% in the English word order, where both of those heads 00:16:25.650 --> 00:16:28.320 align:middle line:84% are red because they're preceding their complements. 00:16:28.320 --> 00:16:30.480 align:middle line:84% You can ask it in the German word 00:16:30.480 --> 00:16:32.310 align:middle line:84% order, where both of those heads are 00:16:32.310 --> 00:16:33.600 align:middle line:90% following their complements. 00:16:33.600 --> 00:16:35.730 align:middle line:84% So you're literally saying, "When Jussi 00:16:35.730 --> 00:16:38.490 align:middle line:90% a novel written would have?" 00:16:38.490 --> 00:16:40.775 align:middle line:84% You can say it in the West Flemish order, where 00:16:40.775 --> 00:16:42.900 align:middle line:84% the auxiliary precedes the verb phrase and the verb 00:16:42.900 --> 00:16:44.250 align:middle line:90% comes after the object. 00:16:44.250 --> 00:16:46.950 align:middle line:84% But you cannot say it in the cross linguistically unattested 00:16:46.950 --> 00:16:48.450 align:middle line:90% order. 00:16:48.450 --> 00:16:50.820 align:middle line:84% So it isn't just that there are no languages like that. 00:16:50.820 --> 00:16:52.530 align:middle line:84% Even in languages like Finnish, where 00:16:52.530 --> 00:16:54.280 align:middle line:84% there's a fair amount of freedom of order, 00:16:54.280 --> 00:16:55.515 align:middle line:90% that order is ruled out. 00:16:55.515 --> 00:16:58.390 align:middle line:90% 00:16:58.390 --> 00:16:59.920 align:middle line:90% Which is weird. 00:16:59.920 --> 00:17:02.080 align:middle line:84% It would be nice to have a theory of that. 00:17:02.080 --> 00:17:06.790 align:middle line:84% People have worked on theories of that. 00:17:06.790 --> 00:17:09.339 align:middle line:84% There's something called the Final-over-Final Constraint 00:17:09.339 --> 00:17:10.180 align:middle line:90% that's been offered. 00:17:10.180 --> 00:17:13.150 align:middle line:84% It's called the FOFC by people who are into it. 00:17:13.150 --> 00:17:17.119 align:middle line:84% What the FOFC says is, at least for certain parts of the tree, 00:17:17.119 --> 00:17:21.730 align:middle line:84% if you have two heads, A and B, and A has, 00:17:21.730 --> 00:17:24.490 align:middle line:84% inside its complement, another head, B, 00:17:24.490 --> 00:17:29.710 align:middle line:84% then if A is head final--- if A follows its complement-- 00:17:29.710 --> 00:17:32.380 align:middle line:84% B also has to follow its complement. 00:17:32.380 --> 00:17:34.930 align:middle line:84% It's as though, as you're building the tree-- 00:17:34.930 --> 00:17:36.760 align:middle line:84% we've been building trees sort of way 00:17:36.760 --> 00:17:38.980 align:middle line:84% we built this one, where you start-- 00:17:38.980 --> 00:17:40.450 align:middle line:84% you're doing repeated merge, right? 00:17:40.450 --> 00:17:41.950 align:middle line:84% You start at the bottom of the tree, 00:17:41.950 --> 00:17:43.780 align:middle line:84% and you keep adding things, and the tree 00:17:43.780 --> 00:17:44.950 align:middle line:90% gets larger and larger. 00:17:44.950 --> 00:17:48.280 align:middle line:84% It's as though, at the beginning, 00:17:48.280 --> 00:17:51.130 align:middle line:84% you decide whether your heads are going to proceed or follow 00:17:51.130 --> 00:17:52.090 align:middle line:90% their complements. 00:17:52.090 --> 00:17:57.430 align:middle line:84% And then you can switch to being head initial, but just once. 00:17:57.430 --> 00:17:59.410 align:middle line:84% After you've decided to be head initial, 00:17:59.410 --> 00:18:01.180 align:middle line:84% you have to be head initial from then on. 00:18:01.180 --> 00:18:04.150 align:middle line:84% You can't go back to being head final, something like that. 00:18:04.150 --> 00:18:05.342 align:middle line:90% You can switch from being-- 00:18:05.342 --> 00:18:07.300 align:middle line:84% as you're building the tree from the bottom up, 00:18:07.300 --> 00:18:09.400 align:middle line:84% you can switch from being head final to being head initial. 00:18:09.400 --> 00:18:11.055 align:middle line:84% You can switch from having heads follow 00:18:11.055 --> 00:18:12.430 align:middle line:84% their complements to having heads 00:18:12.430 --> 00:18:13.540 align:middle line:90% precede their complements. 00:18:13.540 --> 00:18:15.220 align:middle line:84% That's what West Flemish does, where 00:18:15.220 --> 00:18:16.930 align:middle line:84% it has the verb after the object and then 00:18:16.930 --> 00:18:19.570 align:middle line:84% it has the auxiliary before the verb phrase. 00:18:19.570 --> 00:18:21.830 align:middle line:84% But you can't switch in the other direction. 00:18:21.830 --> 00:18:23.080 align:middle line:90% That's what this seems to say. 00:18:23.080 --> 00:18:24.850 align:middle line:84% For certain parts of the tree-- we 00:18:24.850 --> 00:18:26.930 align:middle line:84% may get a chance to talk more about that. 00:18:26.930 --> 00:18:27.430 align:middle line:90% Yeah? 00:18:27.430 --> 00:18:29.013 align:middle line:84% AUDIENCE: What is head and complement? 00:18:29.013 --> 00:18:30.138 align:middle line:90% NORVIN RICHARDS: I'm sorry. 00:18:30.138 --> 00:18:32.170 align:middle line:84% By head and complement, I just mean-- 00:18:32.170 --> 00:18:37.270 align:middle line:84% so heads are things like this, nodes that just dominate a word 00:18:37.270 --> 00:18:39.920 align:middle line:90% and don't contain anything else. 00:18:39.920 --> 00:18:43.330 align:middle line:84% And the complement of the head is its sister. 00:18:43.330 --> 00:18:49.640 align:middle line:84% So here's a head and here's its complement. 00:18:49.640 --> 00:18:54.880 align:middle line:84% So this VP is the complement of this head, or this head has, 00:18:54.880 --> 00:18:56.890 align:middle line:84% as its complement, the "sushi," or this head 00:18:56.890 --> 00:18:59.530 align:middle line:84% has its complement this noun phrase, "the chopsticks." 00:18:59.530 --> 00:19:04.252 align:middle line:84% And the observation is that if you have-- 00:19:04.252 --> 00:19:05.850 align:middle line:84% you can have final heads lower down. 00:19:05.850 --> 00:19:08.350 align:middle line:84% Yeah, you can have heads that follow their complements lower 00:19:08.350 --> 00:19:10.767 align:middle line:84% down and heads that precede their complements higher up, 00:19:10.767 --> 00:19:12.100 align:middle line:90% but you can't have the opposite. 00:19:12.100 --> 00:19:13.036 align:middle line:90% Yeah? 00:19:13.036 --> 00:19:15.458 align:middle line:84% AUDIENCE: So, in this example, we're 00:19:15.458 --> 00:19:18.820 align:middle line:90% saying that we have phonemes. 00:19:18.820 --> 00:19:20.290 align:middle line:90% We don't get to-- 00:19:20.290 --> 00:19:21.786 align:middle line:90% NORVIN RICHARDS: Yup. 00:19:21.786 --> 00:19:24.780 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:19:24.780 --> 00:19:28.478 align:middle line:90% 00:19:28.478 --> 00:19:30.270 align:middle line:84% NORVIN RICHARDS: So yeah, what we're saying 00:19:30.270 --> 00:19:34.230 align:middle line:84% is, if we go back to these trees-- what we're saying is, 00:19:34.230 --> 00:19:39.780 align:middle line:84% in the tree on the far right, because the V precedes 00:19:39.780 --> 00:19:42.030 align:middle line:84% its complement, the higher T must also 00:19:42.030 --> 00:19:43.650 align:middle line:90% precede its complement. 00:19:43.650 --> 00:19:46.770 align:middle line:84% That's what the Final-over-Final Constraint is meant to say. 00:19:46.770 --> 00:19:48.990 align:middle line:84% If the lower head follows the complement, 00:19:48.990 --> 00:19:51.900 align:middle line:84% then it's OK for the higher head to precede or follow. 00:19:51.900 --> 00:19:54.240 align:middle line:84% But if the lower head precedes, then the higher head 00:19:54.240 --> 00:19:55.200 align:middle line:90% must also precede. 00:19:55.200 --> 00:19:56.900 align:middle line:90% It cannot follow. 00:19:56.900 --> 00:19:57.900 align:middle line:90% I hope I wrote it right. 00:19:57.900 --> 00:20:00.710 align:middle line:90% 00:20:00.710 --> 00:20:03.410 align:middle line:84% If A follows its complement, B must also-- wait a minute. 00:20:03.410 --> 00:20:05.540 align:middle line:90% Yeah, yeah, yeah. 00:20:05.540 --> 00:20:10.030 align:middle line:84% A follows its complement, B must also-- yes, yes, yes. 00:20:10.030 --> 00:20:14.530 align:middle line:90% So in the trees-- 00:20:14.530 --> 00:20:17.320 align:middle line:84% in the even-numbered trees where the higher head 00:20:17.320 --> 00:20:19.060 align:middle line:84% follows its complement, what we're 00:20:19.060 --> 00:20:21.230 align:middle line:84% seeing-- where T follows its complement. 00:20:21.230 --> 00:20:23.260 align:middle line:84% So that's the A. What we're seeing 00:20:23.260 --> 00:20:26.380 align:middle line:84% is that in trees like that, the lower head must also 00:20:26.380 --> 00:20:27.340 align:middle line:90% follow its complement. 00:20:27.340 --> 00:20:28.723 align:middle line:90% It can't precede. 00:20:28.723 --> 00:20:30.640 align:middle line:84% So that's what the Final-over-Final constraint 00:20:30.640 --> 00:20:31.920 align:middle line:90% says. 00:20:31.920 --> 00:20:33.960 align:middle line:84% It's an unfortunately-named constraint. 00:20:33.960 --> 00:20:37.560 align:middle line:84% It's misleading, but that's what it-- that's what it does. 00:20:37.560 --> 00:20:38.060 align:middle line:90% OK? 00:20:38.060 --> 00:20:39.727 align:middle line:84% So we get that kind of mixed headedness, 00:20:39.727 --> 00:20:42.520 align:middle line:84% the West Flemish kind, but not the kind on the right. 00:20:42.520 --> 00:20:48.340 align:middle line:90% 00:20:48.340 --> 00:20:54.250 align:middle line:84% Yeah, so-- oh, I mixed up my terminology here. 00:20:54.250 --> 00:20:54.750 align:middle line:90% I'm sorry. 00:20:54.750 --> 00:20:56.550 align:middle line:84% I'll fix this slide before I post it. 00:20:56.550 --> 00:20:58.580 align:middle line:90% So this is the FOFC violation. 00:20:58.580 --> 00:21:00.950 align:middle line:84% T-- I called it I, which I shouldn't have done. 00:21:00.950 --> 00:21:02.510 align:middle line:90% That's an older word for TP. 00:21:02.510 --> 00:21:05.540 align:middle line:84% T has V in its complement, and T follows VP, 00:21:05.540 --> 00:21:07.220 align:middle line:90% but V is not following NP. 00:21:07.220 --> 00:21:08.990 align:middle line:84% That's what makes that a Final-over-Final 00:21:08.990 --> 00:21:10.840 align:middle line:90% Concern violation. 00:21:10.840 --> 00:21:11.340 align:middle line:90% OK? 00:21:11.340 --> 00:21:14.690 align:middle line:90% 00:21:14.690 --> 00:21:16.470 align:middle line:90% For certain parts of the tree. 00:21:16.470 --> 00:21:18.260 align:middle line:84% Now, for certain parts of the tree, 00:21:18.260 --> 00:21:20.360 align:middle line:84% there have to be restrictions on how this applies. 00:21:20.360 --> 00:21:24.530 align:middle line:84% So German, for example, we've just 00:21:24.530 --> 00:21:29.960 align:middle line:84% said that if the higher heads are final, 00:21:29.960 --> 00:21:32.450 align:middle line:84% the lower head also has to be final. 00:21:32.450 --> 00:21:36.260 align:middle line:84% So the fact that the T is final means that the V also 00:21:36.260 --> 00:21:37.250 align:middle line:90% has to be final. 00:21:37.250 --> 00:21:38.450 align:middle line:90% So far, so good. 00:21:38.450 --> 00:21:42.500 align:middle line:84% But the P is head initial, so the people who do the FOFC 00:21:42.500 --> 00:21:44.540 align:middle line:84% have to talk about domains of the tree 00:21:44.540 --> 00:21:46.220 align:middle line:90% in which the FOFC applies. 00:21:46.220 --> 00:21:49.550 align:middle line:84% It applies to the relationship between T and the verb, 00:21:49.550 --> 00:21:51.950 align:middle line:84% but not to the relationship between, say, the verb 00:21:51.950 --> 00:21:53.390 align:middle line:90% and the prepositional phrase-- 00:21:53.390 --> 00:21:54.470 align:middle line:90% prepositional phrases. 00:21:54.470 --> 00:21:58.070 align:middle line:84% It's like they talk as though-- the FOFC kind of starts over 00:21:58.070 --> 00:21:58.970 align:middle line:90% at certain points. 00:21:58.970 --> 00:22:01.648 align:middle line:84% So you calculate it within the prepositional phrase, 00:22:01.648 --> 00:22:03.440 align:middle line:84% or you calculate it within the noun phrase, 00:22:03.440 --> 00:22:05.485 align:middle line:90% but then you start again. 00:22:05.485 --> 00:22:06.860 align:middle line:84% There are other parts of the tree 00:22:06.860 --> 00:22:10.958 align:middle line:84% to which the FOFC applies, but not across those boundaries. 00:22:10.958 --> 00:22:13.250 align:middle line:84% Lots of work on trying to figure out why that would be. 00:22:13.250 --> 00:22:16.190 align:middle line:84% Lots of questions about the FOFC, like why is it true 00:22:16.190 --> 00:22:19.520 align:middle line:84% and which chunks of the tree is it true in? 00:22:19.520 --> 00:22:21.860 align:middle line:84% But I'm telling you about it because it 00:22:21.860 --> 00:22:23.990 align:middle line:84% looks like another case of a linguistic universal, 00:22:23.990 --> 00:22:26.840 align:middle line:84% so another place where, yes, there's 00:22:26.840 --> 00:22:28.370 align:middle line:90% more than one kind of language. 00:22:28.370 --> 00:22:30.650 align:middle line:84% There are head initial languages like English. 00:22:30.650 --> 00:22:33.075 align:middle line:84% There are head final languages like Japanese. 00:22:33.075 --> 00:22:35.700 align:middle line:84% There are languages like German, which have some initial heads, 00:22:35.700 --> 00:22:36.650 align:middle line:90% and other final heads. 00:22:36.650 --> 00:22:39.260 align:middle line:84% But you don't get every logically possible combination 00:22:39.260 --> 00:22:40.730 align:middle line:90% of initial and final. 00:22:40.730 --> 00:22:42.985 align:middle line:90% That's the result of this. 00:22:42.985 --> 00:22:44.610 align:middle line:84% And the FOFC is meant to describe that. 00:22:44.610 --> 00:22:47.027 align:middle line:84% And then, of course, we want to know why the FOFC is true, 00:22:47.027 --> 00:22:50.180 align:middle line:90% if it turns out to be true. 00:22:50.180 --> 00:22:52.398 align:middle line:90% Yeah? 00:22:52.398 --> 00:22:54.158 align:middle line:90% OK. 00:22:54.158 --> 00:22:54.950 align:middle line:90% Does it make sense? 00:22:54.950 --> 00:22:56.325 align:middle line:84% Do we have a question about this? 00:22:56.325 --> 00:22:56.952 align:middle line:90% FOFC? 00:22:56.952 --> 00:22:58.160 align:middle line:90% I'm sorry, I called you FOFC. 00:22:58.160 --> 00:22:58.660 align:middle line:90% Faith? 00:22:58.660 --> 00:23:01.130 align:middle line:90% 00:23:01.130 --> 00:23:03.160 align:middle line:90% I will stop calling you FOFC. 00:23:03.160 --> 00:23:06.020 align:middle line:84% AUDIENCE: Just a question about linguistic universals. 00:23:06.020 --> 00:23:08.587 align:middle line:84% So is it always that something doesn't exist? 00:23:08.587 --> 00:23:10.170 align:middle line:84% Or are there any linguistic universals 00:23:10.170 --> 00:23:13.130 align:middle line:84% where we can say, like, oh, this is the case in every language? 00:23:13.130 --> 00:23:15.800 align:middle line:90% NORVIN RICHARDS: Oh, oh, oh. 00:23:15.800 --> 00:23:18.230 align:middle line:84% I mean, sometimes this is just a matter 00:23:18.230 --> 00:23:20.550 align:middle line:84% of stating the universals in a particular way. 00:23:20.550 --> 00:23:25.070 align:middle line:84% So every language has the property 00:23:25.070 --> 00:23:26.390 align:middle line:90% that it obeys the FOFC. 00:23:26.390 --> 00:23:28.262 align:middle line:90% There. 00:23:28.262 --> 00:23:30.180 align:middle line:84% There are some other things sort of like that. 00:23:30.180 --> 00:23:33.680 align:middle line:84% So for example, it is true of every language 00:23:33.680 --> 00:23:36.560 align:middle line:84% in which the verb comes at the beginning of the clause 00:23:36.560 --> 00:23:41.600 align:middle line:84% that wh-movement happens overtly, at least optionally. 00:23:41.600 --> 00:23:44.960 align:middle line:84% So we said there are languages that do overt wh-movement. 00:23:44.960 --> 00:23:46.760 align:middle line:84% There are languages that have wh-in-situ. 00:23:46.760 --> 00:23:49.530 align:middle line:84% There are languages that have both options. 00:23:49.530 --> 00:23:51.385 align:middle line:84% And there are verb initial languages 00:23:51.385 --> 00:23:52.760 align:middle line:84% that have both options, and there 00:23:52.760 --> 00:23:54.927 align:middle line:84% are verb initial languages that have overt movement. 00:23:54.927 --> 00:23:56.390 align:middle line:84% There are no verb initial languages 00:23:56.390 --> 00:23:58.070 align:middle line:90% with obligatory wh-in-situ. 00:23:58.070 --> 00:23:59.010 align:middle line:90% That never happens. 00:23:59.010 --> 00:24:02.873 align:middle line:84% And again, we want to understand why, but that's a fact. 00:24:02.873 --> 00:24:04.290 align:middle line:84% These parts where I keep saying we 00:24:04.290 --> 00:24:07.860 align:middle line:84% want to understand why, that's because there is work on this. 00:24:07.860 --> 00:24:10.140 align:middle line:84% I have a book that tries to derive that fact I just 00:24:10.140 --> 00:24:11.430 align:middle line:90% told you about. 00:24:11.430 --> 00:24:15.540 align:middle line:84% But I feel-- I would feel bad teaching you 00:24:15.540 --> 00:24:17.760 align:middle line:84% theories that I have posited in a book 00:24:17.760 --> 00:24:19.770 align:middle line:90% because I could be wrong. 00:24:19.770 --> 00:24:22.200 align:middle line:84% I know, it's hard to believe, but I could be. 00:24:22.200 --> 00:24:24.450 align:middle line:84% I'm only trying to tell you things that are definitely 00:24:24.450 --> 00:24:27.000 align:middle line:90% true in this class. 00:24:27.000 --> 00:24:27.700 align:middle line:90% Good question. 00:24:27.700 --> 00:24:30.150 align:middle line:90% Are there other questions? 00:24:30.150 --> 00:24:31.420 align:middle line:90% OK. 00:24:31.420 --> 00:24:37.470 align:middle line:84% All right, so that was it for German. 00:24:37.470 --> 00:24:38.310 align:middle line:90% Oh right, sorry. 00:24:38.310 --> 00:24:40.650 align:middle line:90% Back to German. 00:24:40.650 --> 00:24:42.180 align:middle line:90% OK, so here we are doing German. 00:24:42.180 --> 00:24:46.900 align:middle line:90% Now I said-- 00:24:46.900 --> 00:24:50.460 align:middle line:84% I think I said this accurately, that in this kind of clause, 00:24:50.460 --> 00:24:53.520 align:middle line:84% in German, the verb comes after the object, 00:24:53.520 --> 00:24:56.410 align:middle line:84% and tense comes after the verb phrase. 00:24:56.410 --> 00:24:59.430 align:middle line:84% And I think I saw people who know some German furrow 00:24:59.430 --> 00:25:02.220 align:middle line:84% their brows at me because that's not the word 00:25:02.220 --> 00:25:04.020 align:middle line:90% order in a German main clause. 00:25:04.020 --> 00:25:05.520 align:middle line:84% So I want to talk now about the word 00:25:05.520 --> 00:25:07.980 align:middle line:90% order in a German main clause. 00:25:07.980 --> 00:25:09.960 align:middle line:84% Here's the word order in a German main clause. 00:25:09.960 --> 00:25:13.020 align:middle line:84% You say in German things like "Mary could 00:25:13.020 --> 00:25:15.700 align:middle line:90% with chopsticks sushi eat." 00:25:15.700 --> 00:25:18.420 align:middle line:84% So the verb is still after its object, 00:25:18.420 --> 00:25:20.610 align:middle line:90% but T is no longer at the end. 00:25:20.610 --> 00:25:22.170 align:middle line:90% It's now, well, earlier. 00:25:22.170 --> 00:25:24.870 align:middle line:90% It's right after "Mary." 00:25:24.870 --> 00:25:28.380 align:middle line:84% This is a property of German main clauses 00:25:28.380 --> 00:25:33.420 align:middle line:84% that the thing that's in T, if there is something that's in T, 00:25:33.420 --> 00:25:38.140 align:middle line:84% has to be preceded by exactly one phrase. 00:25:38.140 --> 00:25:41.250 align:middle line:84% So you can say, "Mary could with chopsticks sushi eat." 00:25:41.250 --> 00:25:45.600 align:middle line:84% You can say, "With chopsticks could Mary sushi eat." 00:25:45.600 --> 00:25:49.130 align:middle line:84% You can say "Sushi could Mary with chopsticks eat." 00:25:49.130 --> 00:25:51.810 align:middle line:84% You get to take some phrase and put it first. 00:25:51.810 --> 00:25:55.410 align:middle line:84% And then what's in T-- what would be in T, 00:25:55.410 --> 00:25:58.590 align:middle line:90% "could," that has to be second. 00:25:58.590 --> 00:26:01.320 align:middle line:90% This is called verb second. 00:26:01.320 --> 00:26:03.618 align:middle line:84% These are examples where there's an auxiliary 00:26:03.618 --> 00:26:04.410 align:middle line:90% in second position. 00:26:04.410 --> 00:26:06.077 align:middle line:84% If there is no auxiliary, then the thing 00:26:06.077 --> 00:26:08.590 align:middle line:84% that goes in second position is the verb. 00:26:08.590 --> 00:26:12.630 align:middle line:84% So if you wanted to say "Mary eats sushi with chopsticks," 00:26:12.630 --> 00:26:15.245 align:middle line:84% so there's no auxiliary, you would take the verb "eat." 00:26:15.245 --> 00:26:16.620 align:middle line:84% You'd put it in a different form. 00:26:16.620 --> 00:26:19.380 align:middle line:84% It would be [GERMAN] and it would go in the place 00:26:19.380 --> 00:26:21.330 align:middle line:84% where "could" is going in these sentences. 00:26:21.330 --> 00:26:24.490 align:middle line:84% Something else would be in first position. 00:26:24.490 --> 00:26:26.790 align:middle line:84% So this is a phenomenon called V2. 00:26:26.790 --> 00:26:30.360 align:middle line:84% It's called V2 because you must take exactly one phrase 00:26:30.360 --> 00:26:35.160 align:middle line:84% and put it first, and then T or the verb goes second. 00:26:35.160 --> 00:26:37.440 align:middle line:84% So you can say all of these things in German, 00:26:37.440 --> 00:26:40.625 align:middle line:84% but you cannot say, for example, "With chopsticks sushi could 00:26:40.625 --> 00:26:42.000 align:middle line:84% Mary eat," or anything like that. 00:26:42.000 --> 00:26:44.042 align:middle line:84% There can't be two phrases in the first position. 00:26:44.042 --> 00:26:46.770 align:middle line:90% 00:26:46.770 --> 00:26:51.420 align:middle line:84% There must be exactly one phrase in the first position. 00:26:51.420 --> 00:26:54.260 align:middle line:84% So verb second-- German clauses, German main clauses-- 00:26:54.260 --> 00:26:56.210 align:middle line:84% actually, we'll refine that in just a second. 00:26:56.210 --> 00:26:57.668 align:middle line:84% They have to start with exactly one 00:26:57.668 --> 00:27:00.257 align:middle line:84% phrase followed by the "verb," where 00:27:00.257 --> 00:27:01.340 align:middle line:90% I've put "verb" in quotes. 00:27:01.340 --> 00:27:03.632 align:middle line:84% Because as you've seen, it doesn't have to be the verb. 00:27:03.632 --> 00:27:06.030 align:middle line:84% It's actually whatever tense is on. 00:27:06.030 --> 00:27:08.600 align:middle line:84% So if there's an auxiliary, then that's 00:27:08.600 --> 00:27:10.310 align:middle line:90% the thing that goes first. 00:27:10.310 --> 00:27:13.055 align:middle line:84% If there's no auxiliary, then tense is realized on the verb. 00:27:13.055 --> 00:27:14.990 align:middle line:84% The verb is pronounced with tense morphology, 00:27:14.990 --> 00:27:17.770 align:middle line:84% and that's what goes in second position. 00:27:17.770 --> 00:27:19.020 align:middle line:90% Why am I telling you all this? 00:27:19.020 --> 00:27:21.830 align:middle line:84% Well, partly so that you can be more fully educated people. 00:27:21.830 --> 00:27:25.870 align:middle line:84% Now you know some things about the grammar of German. 00:27:25.870 --> 00:27:28.420 align:middle line:90% But actually-- oh, sorry, yes. 00:27:28.420 --> 00:27:31.690 align:middle line:84% That's the wrong rhetorical move to make. 00:27:31.690 --> 00:27:34.152 align:middle line:90% Sorry, more German. 00:27:34.152 --> 00:27:36.110 align:middle line:84% No, actually, let me finish the rhetorical move 00:27:36.110 --> 00:27:36.730 align:middle line:90% I was just making. 00:27:36.730 --> 00:27:39.010 align:middle line:84% I'm telling you all this about German partly to show you some 00:27:39.010 --> 00:27:39.850 align:middle line:90% cool things about German. 00:27:39.850 --> 00:27:41.308 align:middle line:84% But what we're going to see shortly 00:27:41.308 --> 00:27:42.700 align:middle line:90% is that this is not just German. 00:27:42.700 --> 00:27:45.800 align:middle line:84% V2 is a cross linguistically common phenomenon. 00:27:45.800 --> 00:27:49.330 align:middle line:84% There are many languages out there that are V2. 00:27:49.330 --> 00:27:52.990 align:middle line:84% So learning this fact about-- these facts about German 00:27:52.990 --> 00:27:55.420 align:middle line:84% will do you some good when you are confronted with, 00:27:55.420 --> 00:27:58.810 align:middle line:84% for example, Kashmiri or Dinka or a number of other languages 00:27:58.810 --> 00:28:01.230 align:middle line:84% that are out there which are also V2. 00:28:01.230 --> 00:28:04.382 align:middle line:84% OK, but first, I just said German main clauses, 00:28:04.382 --> 00:28:06.340 align:middle line:84% and then I waffled a little bit about whether I 00:28:06.340 --> 00:28:07.540 align:middle line:90% meant main clauses. 00:28:07.540 --> 00:28:09.730 align:middle line:84% It isn't actually just main clauses. 00:28:09.730 --> 00:28:13.480 align:middle line:84% What it is is clauses that don't have complementizers, 00:28:13.480 --> 00:28:15.400 align:middle line:84% don't have overt complementizers. 00:28:15.400 --> 00:28:20.590 align:middle line:84% So main clauses are V2, but also embedded clauses 00:28:20.590 --> 00:28:22.750 align:middle line:84% as long as they don't start with "dass." 00:28:22.750 --> 00:28:25.990 align:middle line:84% So you can say there are what's called embedded V2, where you 00:28:25.990 --> 00:28:29.120 align:middle line:84% have an embedded clause like the one on the German example 00:28:29.120 --> 00:28:33.730 align:middle line:84% at the bottom of this slide, where you're 00:28:33.730 --> 00:28:35.390 align:middle line:90% doing the V2 phenomenon again. 00:28:35.390 --> 00:28:37.960 align:middle line:84% And I can't remember whether I proved that. 00:28:37.960 --> 00:28:38.860 align:middle line:90% Yeah, so I didn't. 00:28:38.860 --> 00:28:42.010 align:middle line:84% But you could put any phrase in the place of "Mary." 00:28:42.010 --> 00:28:44.250 align:middle line:84% All the shenanigans that we did with the main clause 00:28:44.250 --> 00:28:46.750 align:middle line:84% in the last slide, you can do with the embedded clause here. 00:28:46.750 --> 00:28:50.792 align:middle line:84% So he said, "Mary wanted with chopsticks sushi eat." 00:28:50.792 --> 00:28:52.750 align:middle line:84% You could have put "with chopsticks" or "sushi" 00:28:52.750 --> 00:28:54.130 align:middle line:84% in the place where "Mary" is, and then 00:28:54.130 --> 00:28:56.047 align:middle line:84% "Mary" would have to be after "wanted," right? 00:28:56.047 --> 00:28:58.510 align:middle line:84% So there's one phrase that goes before the auxiliary 00:28:58.510 --> 00:29:03.400 align:middle line:84% in that embedded clause, in case of embedded V2. 00:29:03.400 --> 00:29:05.890 align:middle line:84% What's the difference between this type of embedded clause 00:29:05.890 --> 00:29:09.910 align:middle line:84% and the kinds that we looked at in the first German sentence 00:29:09.910 --> 00:29:12.010 align:middle line:84% when we were just talking about mixed headedness? 00:29:12.010 --> 00:29:14.385 align:middle line:84% Well, it has to do with whether there's a complementizer. 00:29:14.385 --> 00:29:15.850 align:middle line:84% So if your embedded clause starts 00:29:15.850 --> 00:29:18.280 align:middle line:84% with "dass," which is the German word for "that," 00:29:18.280 --> 00:29:20.830 align:middle line:84% the German complementizer for declarative clauses, 00:29:20.830 --> 00:29:22.360 align:middle line:90% then you don't get V2. 00:29:22.360 --> 00:29:25.120 align:middle line:84% And main clauses which don't have a complementizer, 00:29:25.120 --> 00:29:29.320 align:middle line:84% just like in English, also have V2. 00:29:29.320 --> 00:29:31.960 align:middle line:84% So V2 happens whenever there isn't a complementizer being 00:29:31.960 --> 00:29:33.540 align:middle line:90% pronounced. 00:29:33.540 --> 00:29:37.690 align:middle line:84% A hypothesis that people have had about that goes like this. 00:29:37.690 --> 00:29:40.420 align:middle line:84% The tree for German is exactly the tree 00:29:40.420 --> 00:29:41.980 align:middle line:90% that I drew for you before. 00:29:41.980 --> 00:29:45.280 align:middle line:84% It's got a complementizer that precedes its complement. 00:29:45.280 --> 00:29:47.560 align:middle line:84% And then under that complementizer, 00:29:47.560 --> 00:29:51.410 align:middle line:84% everything is head final in the verbal part of the clause. 00:29:51.410 --> 00:29:54.340 align:middle line:84% So the auxiliary follows the verb phrase and the verb 00:29:54.340 --> 00:29:55.510 align:middle line:90% follows its object. 00:29:55.510 --> 00:29:57.670 align:middle line:84% And then, yes, German has prepositions 00:29:57.670 --> 00:30:00.790 align:middle line:84% ahead of postpositions because the FOFC resets itself 00:30:00.790 --> 00:30:03.580 align:middle line:84% across prepositional phrase boundaries. 00:30:03.580 --> 00:30:07.630 align:middle line:84% So that's the tree that we drew before for German clauses. 00:30:07.630 --> 00:30:10.000 align:middle line:84% Complementizers go before the TP. 00:30:10.000 --> 00:30:13.000 align:middle line:84% And then what we're learning is that if nothing is pronounced 00:30:13.000 --> 00:30:16.300 align:middle line:84% in the complimentizer-- if you don't have a "dass"-- 00:30:16.300 --> 00:30:18.990 align:middle line:90% then two things happen. 00:30:18.990 --> 00:30:22.220 align:middle line:84% First, some phrase-- you get to pick some phrase at random-- 00:30:22.220 --> 00:30:25.370 align:middle line:90% moves into the specifier of CP. 00:30:25.370 --> 00:30:27.440 align:middle line:84% So here I've chosen to move the "sushi" 00:30:27.440 --> 00:30:29.480 align:middle line:90% into the specifier of CP. 00:30:29.480 --> 00:30:34.450 align:middle line:84% And also, some head moves in to C. 00:30:34.450 --> 00:30:36.160 align:middle line:84% The consequence of these two movements 00:30:36.160 --> 00:30:38.860 align:middle line:90% is the pattern that we just saw. 00:30:38.860 --> 00:30:41.680 align:middle line:84% First of all, the auxiliary, although we've 00:30:41.680 --> 00:30:43.120 align:middle line:84% seen that when there is a "dass," 00:30:43.120 --> 00:30:44.920 align:middle line:90% the auxiliary is head final. 00:30:44.920 --> 00:30:47.710 align:middle line:84% The auxiliary goes where the complementizer would be. 00:30:47.710 --> 00:30:49.300 align:middle line:90% It becomes head initial. 00:30:49.300 --> 00:30:52.190 align:middle line:84% So it goes almost at the beginning of the sentence, 00:30:52.190 --> 00:30:53.470 align:middle line:90% but not quite. 00:30:53.470 --> 00:30:56.650 align:middle line:84% There has to be exactly one phrase to its left. 00:30:56.650 --> 00:30:58.720 align:middle line:84% And the story is that's the phrase that's 00:30:58.720 --> 00:31:02.240 align:middle line:90% at the left edge of CP. 00:31:02.240 --> 00:31:05.390 align:middle line:84% Raising many questions, like "Why, Germans? 00:31:05.390 --> 00:31:06.930 align:middle line:90% Why are you doing this?" 00:31:06.930 --> 00:31:09.860 align:middle line:84% But as I say, they do, and there are many languages out there 00:31:09.860 --> 00:31:10.525 align:middle line:90% that do this. 00:31:10.525 --> 00:31:11.900 align:middle line:84% There's a lot of interesting work 00:31:11.900 --> 00:31:16.160 align:middle line:84% on how you choose which phrase to randomly move 00:31:16.160 --> 00:31:17.370 align:middle line:90% into first position. 00:31:17.370 --> 00:31:19.468 align:middle line:84% It has consequences for the interpretation 00:31:19.468 --> 00:31:22.010 align:middle line:84% of the sentence which are pretty subtle and difficult to talk 00:31:22.010 --> 00:31:22.510 align:middle line:90% about. 00:31:22.510 --> 00:31:25.500 align:middle line:84% There are people who talk about them, try to figure them out. 00:31:25.500 --> 00:31:28.580 align:middle line:84% They have to do with which phrase 00:31:28.580 --> 00:31:31.670 align:middle line:84% you're trying to draw someone's attention to, or so on. 00:31:31.670 --> 00:31:35.240 align:middle line:84% It's very difficult stuff to talk about seriously. 00:31:35.240 --> 00:31:37.208 align:middle line:84% But since we're not doing semantics yet, 00:31:37.208 --> 00:31:38.750 align:middle line:84% since we're doing syntax, we can just 00:31:38.750 --> 00:31:40.125 align:middle line:84% say this is a thing that happens. 00:31:40.125 --> 00:31:42.140 align:middle line:84% Some phrase goes into first position, 00:31:42.140 --> 00:31:47.970 align:middle line:84% and then whatever's in T goes into second position. 00:31:47.970 --> 00:31:50.810 align:middle line:84% So you are moving a phrase into the specifier of CP 00:31:50.810 --> 00:31:53.210 align:middle line:90% and you are moving a head-- 00:31:53.210 --> 00:31:55.310 align:middle line:84% in particular, the head T is moving 00:31:55.310 --> 00:32:00.875 align:middle line:84% into C. That's how you get the German V2 order. 00:32:00.875 --> 00:32:01.375 align:middle line:90% OK? 00:32:01.375 --> 00:32:05.220 align:middle line:90% 00:32:05.220 --> 00:32:07.320 align:middle line:90% You're stunned by German. 00:32:07.320 --> 00:32:08.310 align:middle line:90% It's understandable. 00:32:08.310 --> 00:32:09.185 align:middle line:90% It's pretty stunning. 00:32:09.185 --> 00:32:12.090 align:middle line:90% 00:32:12.090 --> 00:32:14.550 align:middle line:84% Now this is the part that I thought 00:32:14.550 --> 00:32:17.640 align:middle line:84% I was going to tell you before, but now I will. 00:32:17.640 --> 00:32:18.630 align:middle line:90% It's not just German. 00:32:18.630 --> 00:32:22.590 align:middle line:84% So there are languages, some of them 00:32:22.590 --> 00:32:24.600 align:middle line:84% completely unrelated to German, others 00:32:24.600 --> 00:32:26.520 align:middle line:84% only very distantly related to German, 00:32:26.520 --> 00:32:28.980 align:middle line:90% that have exactly this setup. 00:32:28.980 --> 00:32:34.770 align:middle line:84% So Kashmiri is an Indo-Aryan language spoken in Kashmir 00:32:34.770 --> 00:32:39.730 align:middle line:84% in which there must be some phrase in first position. 00:32:39.730 --> 00:32:43.530 align:middle line:84% And so you say things like "Ram gave Sham a book." 00:32:43.530 --> 00:32:50.220 align:middle line:84% Some phrase has to go first, and when there's no auxiliary, 00:32:50.220 --> 00:32:52.170 align:middle line:84% the verb goes right after that first phrase. 00:32:52.170 --> 00:32:54.820 align:middle line:84% But when there is an auxiliary, like in the second example, 00:32:54.820 --> 00:32:56.700 align:middle line:84% the auxiliary goes in that second position. 00:32:56.700 --> 00:32:59.760 align:middle line:84% And the verb, in this case, goes at the end. 00:32:59.760 --> 00:33:01.470 align:middle line:84% And the thing that goes in first position 00:33:01.470 --> 00:33:02.850 align:middle line:90% doesn't have to be the subject. 00:33:02.850 --> 00:33:06.840 align:middle line:84% You can see that in the last example, where you've put 00:33:06.840 --> 00:33:08.380 align:middle line:90% an adverb in first position. 00:33:08.380 --> 00:33:11.862 align:middle line:84% So Kashmiri is just a very, very odd dialect of German, 00:33:11.862 --> 00:33:12.570 align:middle line:90% basically, right? 00:33:12.570 --> 00:33:14.670 align:middle line:84% So the word order for-- the rules 00:33:14.670 --> 00:33:16.740 align:middle line:84% for word order for Kashmiri resemble 00:33:16.740 --> 00:33:21.120 align:middle line:84% the rules for word order for German to a surprising degree. 00:33:21.120 --> 00:33:23.760 align:middle line:84% I was making a joke when I said that it's a dialect of German. 00:33:23.760 --> 00:33:24.300 align:middle line:90% It's not. 00:33:24.300 --> 00:33:26.290 align:middle line:84% They're both Indo-European languages, 00:33:26.290 --> 00:33:28.603 align:middle line:84% so they're related, but extremely distantly. 00:33:28.603 --> 00:33:30.270 align:middle line:84% They don't have anything else in common, 00:33:30.270 --> 00:33:33.540 align:middle line:90% but they have this in common. 00:33:33.540 --> 00:33:36.600 align:middle line:84% Vata, which is a Kru language of Ivory Coast, 00:33:36.600 --> 00:33:40.380 align:middle line:84% literally not related to German, also a V2 language. 00:33:40.380 --> 00:33:41.680 align:middle line:90% And a bunch of others-- 00:33:41.680 --> 00:33:44.790 align:middle line:84% Karitiana, which is a Tupian language spoken in Brazil. 00:33:44.790 --> 00:33:47.400 align:middle line:84% Ingush, which is a Nakh-Daghestanian language 00:33:47.400 --> 00:33:49.560 align:middle line:84% spoken in the Caucasus, Dinka, which 00:33:49.560 --> 00:33:52.840 align:middle line:84% is an Nilontic language spoken in South Sudan, V2 is a thing. 00:33:52.840 --> 00:33:54.750 align:middle line:84% It's not a hugely common thing, but it's 00:33:54.750 --> 00:33:56.730 align:middle line:84% a thing you find scattered all over the world. 00:33:56.730 --> 00:33:58.600 align:middle line:84% There are lots of V2 languages out there. 00:33:58.600 --> 00:34:01.020 align:middle line:84% That's one of the factory options for language word 00:34:01.020 --> 00:34:03.390 align:middle line:90% orders, being V2. 00:34:03.390 --> 00:34:04.680 align:middle line:90% Lots of languages out there. 00:34:04.680 --> 00:34:08.489 align:middle line:90% 00:34:08.489 --> 00:34:10.550 align:middle line:84% So V2, the verb second, is all over the place. 00:34:10.550 --> 00:34:12.800 align:middle line:84% And in our system, what it consists of 00:34:12.800 --> 00:34:15.770 align:middle line:84% is head movement to head initial C and movement 00:34:15.770 --> 00:34:18.110 align:middle line:84% of some phrase, some randomly chosen phrase, 00:34:18.110 --> 00:34:20.260 align:middle line:90% to the specifier of CP. 00:34:20.260 --> 00:34:24.830 align:middle line:84% OK. now why am I making such a big deal about this? 00:34:24.830 --> 00:34:28.070 align:middle line:84% We know that complementizer can be final. 00:34:28.070 --> 00:34:31.250 align:middle line:84% So German complementizers are head initial. 00:34:31.250 --> 00:34:33.679 align:middle line:84% We know that by looking at them when they're overt. 00:34:33.679 --> 00:34:35.719 align:middle line:84% So German has this complementizer, "dass," which 00:34:35.719 --> 00:34:36.844 align:middle line:90% is their version of "that." 00:34:36.844 --> 00:34:40.010 align:middle line:84% It's actually related to the English word "that." 00:34:40.010 --> 00:34:41.870 align:middle line:84% We can see that it is head initial. 00:34:41.870 --> 00:34:46.040 align:middle line:84% And we can also see from German V2 that German V2 involves 00:34:46.040 --> 00:34:49.639 align:middle line:84% moving the auxiliary-- or if there is no auxiliary, 00:34:49.639 --> 00:34:50.780 align:middle line:90% the verb-- 00:34:50.780 --> 00:34:54.030 align:middle line:84% into C, which causes it to be in second position. 00:34:54.030 --> 00:34:56.270 align:middle line:84% So now it precedes the verb phrase, 00:34:56.270 --> 00:34:58.520 align:middle line:84% precedes the rest of the clause, and there's 00:34:58.520 --> 00:35:00.440 align:middle line:90% some phrase in its specifier. 00:35:00.440 --> 00:35:01.940 align:middle line:84% We know that there can be languages 00:35:01.940 --> 00:35:02.990 align:middle line:90% in which C is head final. 00:35:02.990 --> 00:35:04.160 align:middle line:90% I showed you that earlier. 00:35:04.160 --> 00:35:05.535 align:middle line:84% There are languages like Japanese 00:35:05.535 --> 00:35:07.340 align:middle line:90% in which C goes at the end. 00:35:07.340 --> 00:35:09.043 align:middle line:84% Japanese has a word for that too. 00:35:09.043 --> 00:35:09.710 align:middle line:90% It's [JAPANESE]. 00:35:09.710 --> 00:35:13.340 align:middle line:84% It goes after the clause that it combines with. 00:35:13.340 --> 00:35:15.950 align:middle line:84% So it wouldn't be hard to build the language that 00:35:15.950 --> 00:35:18.560 align:middle line:84% would be the mirror image of German, right? 00:35:18.560 --> 00:35:21.380 align:middle line:84% So this would be a language in which you'd 00:35:21.380 --> 00:35:24.950 align:middle line:84% have your clause, your TP, and it would have in it 00:35:24.950 --> 00:35:28.310 align:middle line:84% whatever it would have, and there would be a head final C. 00:35:28.310 --> 00:35:37.178 align:middle line:84% And you would move some phrase, XP, into a specifier of CP that 00:35:37.178 --> 00:35:37.970 align:middle line:90% would be over here. 00:35:37.970 --> 00:35:39.680 align:middle line:90% This wouldn't be hard, right? 00:35:39.680 --> 00:35:44.663 align:middle line:84% This is just German, except that in German, the C is over here 00:35:44.663 --> 00:35:46.580 align:middle line:84% and the complementizer that you're moving into 00:35:46.580 --> 00:35:48.450 align:middle line:90% is over here. 00:35:48.450 --> 00:35:49.640 align:middle line:90% So this is German. 00:35:49.640 --> 00:35:52.340 align:middle line:84% There is some phrase in TP which is 00:35:52.340 --> 00:35:54.740 align:middle line:84% moving into the specifier of CP, which is over here. 00:35:54.740 --> 00:35:56.687 align:middle line:90% C is over here, and this is V2. 00:35:56.687 --> 00:35:57.770 align:middle line:90% This is the V2 word order. 00:35:57.770 --> 00:36:01.507 align:middle line:84% This is German and Kashmiri and Kru and Dinka and, and, and. 00:36:01.507 --> 00:36:03.090 align:middle line:84% There are lots of languages like that. 00:36:03.090 --> 00:36:05.940 align:middle line:84% So are there any languages like this? 00:36:05.940 --> 00:36:07.560 align:middle line:90% No. 00:36:07.560 --> 00:36:08.700 align:middle line:90% None. 00:36:08.700 --> 00:36:11.280 align:middle line:84% Lots of languages like this, no languages like that. 00:36:11.280 --> 00:36:14.290 align:middle line:90% That's not a thing. 00:36:14.290 --> 00:36:16.780 align:middle line:84% Raising lots of questions like, well, wait. 00:36:16.780 --> 00:36:18.100 align:middle line:90% Why is that? 00:36:18.100 --> 00:36:19.450 align:middle line:90% Because it's not-- right? 00:36:19.450 --> 00:36:22.960 align:middle line:84% It's not hard to describe that language on the left. 00:36:22.960 --> 00:36:25.782 align:middle line:84% But it doesn't-- sorry, that language on the right. 00:36:25.782 --> 00:36:27.490 align:middle line:84% The language on the left is quite common, 00:36:27.490 --> 00:36:29.653 align:middle line:84% cross linguistically well attested. 00:36:29.653 --> 00:36:31.195 align:middle line:84% Language on the right doesn't happen. 00:36:31.195 --> 00:36:33.920 align:middle line:90% 00:36:33.920 --> 00:36:36.220 align:middle line:84% So could there be a language like this? 00:36:36.220 --> 00:36:40.195 align:middle line:84% No such language has ever been found. 00:36:40.195 --> 00:36:42.070 align:middle line:84% And although there are verb second languages, 00:36:42.070 --> 00:36:44.610 align:middle line:90% plenty of them-- 00:36:44.610 --> 00:36:46.340 align:middle line:84% and so forget about trees for a second. 00:36:46.340 --> 00:36:48.380 align:middle line:84% Suppose you thought-- all of these trees 00:36:48.380 --> 00:36:50.545 align:middle line:84% I've been showing you, put them out of your mind. 00:36:50.545 --> 00:36:51.920 align:middle line:84% We should really just be thinking 00:36:51.920 --> 00:36:55.080 align:middle line:84% about verbs and subjects and objects and things like that. 00:36:55.080 --> 00:36:57.350 align:middle line:90% We should forget about trees. 00:36:57.350 --> 00:36:59.540 align:middle line:84% It's a fact that there are verb second languages. 00:36:59.540 --> 00:37:01.707 align:middle line:84% Again, plenty of them, scattered all over the world, 00:37:01.707 --> 00:37:04.100 align:middle line:84% not related to each other, not the result of contact. 00:37:04.100 --> 00:37:05.900 align:middle line:84% Yeah, it's apparently one of the factory 00:37:05.900 --> 00:37:10.380 align:middle line:84% options for having a human language, is to have it V2. 00:37:10.380 --> 00:37:12.150 align:middle line:84% So you might imagine that there could also 00:37:12.150 --> 00:37:15.360 align:middle line:84% be direct object second languages, 00:37:15.360 --> 00:37:16.590 align:middle line:90% or subject second language. 00:37:16.590 --> 00:37:19.370 align:middle line:84% Languages where the subject had to be the second thing 00:37:19.370 --> 00:37:21.870 align:middle line:84% and you could put anything you wanted to before the subject, 00:37:21.870 --> 00:37:22.380 align:middle line:90% right? 00:37:22.380 --> 00:37:23.940 align:middle line:84% Or languages where the direct object 00:37:23.940 --> 00:37:24.960 align:middle line:84% had to be the second thing, and you 00:37:24.960 --> 00:37:27.750 align:middle line:84% could put anything you wanted to before the direct object. 00:37:27.750 --> 00:37:30.570 align:middle line:84% I wouldn't know how to draw a tree for a language like that, 00:37:30.570 --> 00:37:32.970 align:middle line:84% but it's not any harder to describe it 00:37:32.970 --> 00:37:35.700 align:middle line:84% in words than it is to describe a verb second language. 00:37:35.700 --> 00:37:37.140 align:middle line:90% Do people see what I mean? 00:37:37.140 --> 00:37:39.090 align:middle line:84% This is a reason to take trees seriously 00:37:39.090 --> 00:37:40.650 align:middle line:84% as a way of describing languages, 00:37:40.650 --> 00:37:43.320 align:middle line:84% because they make it easy to describe German-- 00:37:43.320 --> 00:37:46.470 align:middle line:84% we just did it-- but hard to describe these other imaginable 00:37:46.470 --> 00:37:48.360 align:middle line:84% languages, direct object second languages 00:37:48.360 --> 00:37:49.920 align:middle line:90% or subject second languages. 00:37:49.920 --> 00:37:52.810 align:middle line:84% And the fact is that those languages don't exist. 00:37:52.810 --> 00:37:55.560 align:middle line:84% So the German kind of language, the verb second kind 00:37:55.560 --> 00:37:58.802 align:middle line:84% of language, reasonably common in the world. 00:37:58.802 --> 00:38:01.260 align:middle line:84% But these other languages that are not too hard to imagine, 00:38:01.260 --> 00:38:03.640 align:middle line:90% they don't happen. 00:38:03.640 --> 00:38:06.930 align:middle line:84% So we're at another one of these frustrating points in the class 00:38:06.930 --> 00:38:10.200 align:middle line:90% where I've shown you a mystery. 00:38:10.200 --> 00:38:11.550 align:middle line:90% There are verb second languages. 00:38:11.550 --> 00:38:14.160 align:middle line:84% There are no verb penultimate languages, mirror 00:38:14.160 --> 00:38:15.152 align:middle line:90% image of German. 00:38:15.152 --> 00:38:17.610 align:middle line:84% There are verb second languages, there are no direct object 00:38:17.610 --> 00:38:19.680 align:middle line:90% second languages. 00:38:19.680 --> 00:38:22.920 align:middle line:84% And when I showed you the FOFC, we were at a similar point. 00:38:22.920 --> 00:38:24.720 align:middle line:84% The FOFC, it seems to happen, yeah. 00:38:24.720 --> 00:38:27.900 align:middle line:84% So there are languages like English and Japanese 00:38:27.900 --> 00:38:29.190 align:middle line:90% and German and West Flemish. 00:38:29.190 --> 00:38:31.860 align:middle line:84% There's another logically possible kind of language 00:38:31.860 --> 00:38:33.897 align:middle line:84% where the verb phrase is head final and the TP-- 00:38:33.897 --> 00:38:35.730 align:middle line:84% sorry, where the verb phrase is head initial 00:38:35.730 --> 00:38:37.020 align:middle line:90% and the TP is head final. 00:38:37.020 --> 00:38:38.460 align:middle line:90% That language doesn't exist. 00:38:38.460 --> 00:38:40.860 align:middle line:84% And in languages like Finnish and Basque-- 00:38:40.860 --> 00:38:42.540 align:middle line:84% actually, I didn't show you Basque-- 00:38:42.540 --> 00:38:45.930 align:middle line:84% where word order is fairly free under certain circumstances, 00:38:45.930 --> 00:38:48.600 align:middle line:84% that particular word order is ruled out. 00:38:48.600 --> 00:38:52.050 align:middle line:84% And then I said, sure would be nice to know why. 00:38:52.050 --> 00:38:55.295 align:middle line:84% And if this were an intro to syntax class, 00:38:55.295 --> 00:38:57.420 align:middle line:84% this is where I would start trying to show you why, 00:38:57.420 --> 00:38:59.850 align:middle line:84% or showing you people's theories of why. 00:38:59.850 --> 00:39:03.355 align:middle line:84% But instead, this is 24.900, and we've already spent many days 00:39:03.355 --> 00:39:05.730 align:middle line:84% talking about syntax, and I really have to start teaching 00:39:05.730 --> 00:39:06.600 align:middle line:90% you other things soon. 00:39:06.600 --> 00:39:08.100 align:middle line:84% So I will just tell you, if you would 00:39:08.100 --> 00:39:09.933 align:middle line:84% like to know the answers to these questions, 00:39:09.933 --> 00:39:11.308 align:middle line:84% go take more linguistics classes. 00:39:11.308 --> 00:39:12.766 align:middle line:84% Yeah, these are the kinds of things 00:39:12.766 --> 00:39:14.790 align:middle line:84% linguists talk about and try to figure out. 00:39:14.790 --> 00:39:20.470 align:middle line:90% 00:39:20.470 --> 00:39:23.780 align:middle line:90% So, moral of all this-- 00:39:23.780 --> 00:39:26.210 align:middle line:84% there are different kinds of languages in the world, 00:39:26.210 --> 00:39:29.600 align:middle line:84% but the languages that we find in the world 00:39:29.600 --> 00:39:32.540 align:middle line:84% differ in ways that are constrained. 00:39:32.540 --> 00:39:35.900 align:middle line:84% So we don't find every imaginable kind of language. 00:39:35.900 --> 00:39:37.880 align:middle line:84% There are gaps-- sort of interesting gaps. 00:39:37.880 --> 00:39:40.820 align:middle line:84% Sometimes, gaps that you can define 00:39:40.820 --> 00:39:44.450 align:middle line:84% the boundaries of, like with the FOFC, where you say, yeah, 00:39:44.450 --> 00:39:47.060 align:middle line:84% languages get to decide for particular heads 00:39:47.060 --> 00:39:48.980 align:middle line:84% whether they proceed or follow their sisters, 00:39:48.980 --> 00:39:51.940 align:middle line:84% but there are certain patterns that you don't find, 00:39:51.940 --> 00:39:55.166 align:middle line:84% and we want to have a theory of why. 00:39:55.166 --> 00:39:57.050 align:middle line:84% Wh-movements-- there are languages 00:39:57.050 --> 00:39:58.100 align:middle line:90% that have wh-movement. 00:39:58.100 --> 00:39:59.520 align:middle line:84% There are languages that lack it. 00:39:59.520 --> 00:40:01.490 align:middle line:84% But if you have it, it goes to the left. 00:40:01.490 --> 00:40:02.960 align:middle line:90% It never goes to the right. 00:40:02.960 --> 00:40:05.090 align:middle line:84% There are languages that have V2. 00:40:05.090 --> 00:40:07.040 align:middle line:84% There are languages that don't have V2. 00:40:07.040 --> 00:40:10.535 align:middle line:84% But there aren't languages that have V next 00:40:10.535 --> 00:40:14.030 align:middle line:90% to last, a mirror image of V2. 00:40:14.030 --> 00:40:16.640 align:middle line:84% There is variation among languages, 00:40:16.640 --> 00:40:19.250 align:middle line:84% but there are kinds of things that you don't ever find, 00:40:19.250 --> 00:40:22.472 align:middle line:84% certain peculiar patterns that resolutely fail to show up. 00:40:22.472 --> 00:40:24.680 align:middle line:84% This is the kind of thing we're talking about when we 00:40:24.680 --> 00:40:26.240 align:middle line:90% talk about universal grammar. 00:40:26.240 --> 00:40:28.550 align:middle line:84% So that's a phrase you may hear people talk about. 00:40:28.550 --> 00:40:31.080 align:middle line:84% Sometimes you hear it used as a term of abuse. 00:40:31.080 --> 00:40:33.373 align:middle line:84% So there are people who think that it's a dumb idea, 00:40:33.373 --> 00:40:34.790 align:middle line:84% you know, that it's something that 00:40:34.790 --> 00:40:37.782 align:middle line:84% linguists care about for some reason, but nobody else should. 00:40:37.782 --> 00:40:39.740 align:middle line:84% But this is what we mean when we talk about it. 00:40:39.740 --> 00:40:43.340 align:middle line:84% We mean it looks as though part of being a human being 00:40:43.340 --> 00:40:45.920 align:middle line:84% is having the kind of mind that can build language 00:40:45.920 --> 00:40:47.880 align:middle line:90% in some ways, but not others. 00:40:47.880 --> 00:40:49.810 align:middle line:84% And that's what universal grammar is. 00:40:49.810 --> 00:40:51.560 align:middle line:84% It's kind of a bad name for it, because it 00:40:51.560 --> 00:40:57.020 align:middle line:84% sounds like we all start out with the grammar of, I 00:40:57.020 --> 00:40:58.370 align:middle line:90% don't know, Basque, you know? 00:40:58.370 --> 00:41:00.140 align:middle line:84% And then we learn our native languages 00:41:00.140 --> 00:41:02.780 align:middle line:84% by manipulating the universal grammar that we all start with. 00:41:02.780 --> 00:41:03.860 align:middle line:90% That's not the idea. 00:41:03.860 --> 00:41:06.470 align:middle line:84% The idea is, we're constrained in the kinds 00:41:06.470 --> 00:41:09.180 align:middle line:84% of linguistic hypotheses that we can have. 00:41:09.180 --> 00:41:13.280 align:middle line:84% And so we come preloaded with some instructions about how 00:41:13.280 --> 00:41:17.000 align:middle line:84% to build language, and there are certain kinds 00:41:17.000 --> 00:41:19.845 align:middle line:84% of options that we know better than to even consider 00:41:19.845 --> 00:41:21.470 align:middle line:84% when we're learning our first language. 00:41:21.470 --> 00:41:26.090 align:middle line:84% That's what universal grammar is meant to claim. 00:41:26.090 --> 00:41:27.666 align:middle line:90% Questions about any of that? 00:41:27.666 --> 00:41:29.860 align:middle line:90% Does any of this make sense? 00:41:29.860 --> 00:41:32.990 align:middle line:84% This is meant to inoculate you against-- 00:41:32.990 --> 00:41:36.760 align:middle line:84% so in popular science, you will sometimes 00:41:36.760 --> 00:41:38.980 align:middle line:84% read people claiming that universal grammar can't 00:41:38.980 --> 00:41:41.590 align:middle line:84% possibly be true or it's dead or it's been disproven 00:41:41.590 --> 00:41:42.230 align:middle line:90% or whatever. 00:41:42.230 --> 00:41:44.230 align:middle line:90% And this is all hooey. 00:41:44.230 --> 00:41:48.260 align:middle line:84% You should throw away any papers that seem to claim that. 00:41:48.260 --> 00:41:50.915 align:middle line:90% OK? 00:41:50.915 --> 00:41:53.110 align:middle line:90% All right. 00:41:53.110 --> 00:41:55.570 align:middle line:84% So, yeah, in this particular case, no V-penultimate, 00:41:55.570 --> 00:41:58.000 align:middle line:90% no wh-movement to the right. 00:41:58.000 --> 00:42:00.833 align:middle line:84% It's as though maybe heads can either proceed or follow 00:42:00.833 --> 00:42:02.500 align:middle line:84% their sisters with certain restrictions, 00:42:02.500 --> 00:42:05.380 align:middle line:84% as we've seen in the FOFC, but specifiers maybe always 00:42:05.380 --> 00:42:06.430 align:middle line:90% proceed their sisters. 00:42:06.430 --> 00:42:09.850 align:middle line:84% And again, we'd want to know why. 00:42:09.850 --> 00:42:11.560 align:middle line:90% OK. 00:42:11.560 --> 00:42:16.270 align:middle line:84% So we've now-- so here I am pivoting. 00:42:16.270 --> 00:42:20.150 align:middle line:84% You can tell because the font size is changing. 00:42:20.150 --> 00:42:23.780 align:middle line:84% We've now seen several different kinds of movement operations. 00:42:23.780 --> 00:42:25.990 align:middle line:84% So we've talked about the wh-movement, 00:42:25.990 --> 00:42:28.240 align:middle line:84% which is what makes questions like "What did you buy?" 00:42:28.240 --> 00:42:31.090 align:middle line:84% where "what"-- wh moves to the specifier of CP, 00:42:31.090 --> 00:42:33.940 align:middle line:84% which is on the left periphery of the clause. 00:42:33.940 --> 00:42:36.250 align:middle line:84% We've now talked about head movement. 00:42:36.250 --> 00:42:38.470 align:middle line:84% That was part of talking about German V2. 00:42:38.470 --> 00:42:42.820 align:middle line:84% So in V2, which handily I still have a tree for, 00:42:42.820 --> 00:42:45.130 align:middle line:84% some phrase moves to the specifier of CP, 00:42:45.130 --> 00:42:48.940 align:middle line:84% and T moves into C. And that's why 00:42:48.940 --> 00:42:52.840 align:middle line:84% you get word orders like a prepositional phrase, 00:42:52.840 --> 00:43:01.110 align:middle line:84% "With chopsticks will she sushi eat," 00:43:01.110 --> 00:43:03.510 align:middle line:84% which is how you would say that in German. 00:43:03.510 --> 00:43:05.730 align:middle line:84% So you'd have some phrase in first position, 00:43:05.730 --> 00:43:08.720 align:middle line:84% and then whatever it is that bears tense-- 00:43:08.720 --> 00:43:10.440 align:middle line:84% so the highest auxiliary, if there's 00:43:10.440 --> 00:43:13.410 align:middle line:84% an auxiliary, or the verb, if there's no auxiliary-- ends up 00:43:13.410 --> 00:43:16.200 align:middle line:84% in C. And I just casually said, "Hey, look, 00:43:16.200 --> 00:43:17.910 align:middle line:84% this head is moving to this other head." 00:43:17.910 --> 00:43:20.577 align:middle line:90% That's called head movement. 00:43:20.577 --> 00:43:22.410 align:middle line:84% I do a lot of head movement as I'm teaching. 00:43:22.410 --> 00:43:24.960 align:middle line:90% Some of you may have noticed. 00:43:24.960 --> 00:43:27.330 align:middle line:84% And then we've also talked about NP-movement. 00:43:27.330 --> 00:43:29.460 align:middle line:84% That was the case where you took a noun phrase 00:43:29.460 --> 00:43:31.740 align:middle line:84% and moved it into the specifier of TP. 00:43:31.740 --> 00:43:34.980 align:middle line:90% So three kinds of movement then. 00:43:34.980 --> 00:43:39.375 align:middle line:90% We've got "What did you buy?" 00:43:39.375 --> 00:43:42.240 align:middle line:90% 00:43:42.240 --> 00:43:45.510 align:middle line:90% where "what" is wh-moving. 00:43:45.510 --> 00:43:48.460 align:middle line:90% 00:43:48.460 --> 00:43:51.810 align:middle line:90% We've got German. 00:43:51.810 --> 00:44:01.710 align:middle line:84% So in German, "With chopsticks will 00:44:01.710 --> 00:44:05.570 align:middle line:84% she sushi eat," where we have reasons 00:44:05.570 --> 00:44:07.320 align:middle line:84% to think that auxiliaries start at the end 00:44:07.320 --> 00:44:08.910 align:middle line:90% of the cause and head move. 00:44:08.910 --> 00:44:12.090 align:middle line:90% So here's head movement. 00:44:12.090 --> 00:44:14.430 align:middle line:84% And then the last one is NP-movement, 00:44:14.430 --> 00:44:17.310 align:middle line:84% and that was how we were going to deal with sentences 00:44:17.310 --> 00:44:24.840 align:middle line:84% like "The sushi was devoured," where 00:44:24.840 --> 00:44:27.510 align:middle line:84% we think that "devour" selects for an object, 00:44:27.510 --> 00:44:28.770 align:middle line:90% needs to have an object. 00:44:28.770 --> 00:44:30.690 align:middle line:84% But in this case, its object isn't there 00:44:30.690 --> 00:44:31.980 align:middle line:90% because it's moved over here. 00:44:31.980 --> 00:44:33.960 align:middle line:90% This is NP-movement. 00:44:33.960 --> 00:44:35.040 align:middle line:90% It's moved over here. 00:44:35.040 --> 00:44:39.120 align:middle line:84% Because of the EPP, TP needs to have a specifier, 00:44:39.120 --> 00:44:41.970 align:middle line:84% and the sushi has raised to become the specifier of TP. 00:44:41.970 --> 00:44:45.570 align:middle line:84% So wh-movement movement and head movement and NP-movement, and I 00:44:45.570 --> 00:44:47.730 align:middle line:84% can see that I've used my own private abbreviation 00:44:47.730 --> 00:44:50.490 align:middle line:90% for movements, which is mvmt. 00:44:50.490 --> 00:44:52.107 align:middle line:84% Because if you're a syntactician, 00:44:52.107 --> 00:44:53.940 align:middle line:84% you have to write the word "movement" a lot. 00:44:53.940 --> 00:44:56.850 align:middle line:84% So that's my abbreviation for movement. 00:44:56.850 --> 00:44:59.517 align:middle line:84% Yeah, so three kinds of movement that we've talked about so far. 00:44:59.517 --> 00:45:01.267 align:middle line:84% Are there any questions about any of that? 00:45:01.267 --> 00:45:03.690 align:middle line:84% Because we're about to talk about those movements 00:45:03.690 --> 00:45:05.670 align:middle line:84% and their properties in a little bit of detail. 00:45:05.670 --> 00:45:08.467 align:middle line:90% 00:45:08.467 --> 00:45:08.967 align:middle line:90% OK. 00:45:08.967 --> 00:45:12.270 align:middle line:90% 00:45:12.270 --> 00:45:15.420 align:middle line:90% So "Mary will type novels." 00:45:15.420 --> 00:45:17.693 align:middle line:84% We've talked about head movements. 00:45:17.693 --> 00:45:19.360 align:middle line:84% We haven't talked about it in this case, 00:45:19.360 --> 00:45:21.360 align:middle line:84% but here's a place where we could talk about it. 00:45:21.360 --> 00:45:24.520 align:middle line:84% There's a way to ask questions, yes/no questions, where 00:45:24.520 --> 00:45:27.100 align:middle line:84% you take whatever's in T and you move it into C. 00:45:27.100 --> 00:45:29.200 align:middle line:84% So you start with "Mary will type novels." 00:45:29.200 --> 00:45:32.830 align:middle line:84% You move T into C, and so you have the word order "Will Mary 00:45:32.830 --> 00:45:33.520 align:middle line:90% type novels?" 00:45:33.520 --> 00:45:36.280 align:middle line:84% So the auxiliary now precedes the subject. 00:45:36.280 --> 00:45:38.710 align:middle line:84% So here's an instance of head movement-- in this case, 00:45:38.710 --> 00:45:40.720 align:middle line:84% in order to form yes/no questions in English. 00:45:40.720 --> 00:45:43.540 align:middle line:90% 00:45:43.540 --> 00:45:46.850 align:middle line:84% Wh-questions, "What what will Mary type?" 00:45:46.850 --> 00:45:49.250 align:middle line:84% Not only are you moving T into C-- 00:45:49.250 --> 00:45:51.540 align:middle line:84% so "will" is moving from the position after "Mary" 00:45:51.540 --> 00:45:53.540 align:middle line:84% to the position before "Mary"-- but you're doing 00:45:53.540 --> 00:45:56.060 align:middle line:84% the wh-movement that we've talked about before. 00:45:56.060 --> 00:45:59.840 align:middle line:84% The object is moving into the specifier of CP. 00:45:59.840 --> 00:46:03.493 align:middle line:84% And then in NP-movement, "The cookies were devoured." 00:46:03.493 --> 00:46:04.910 align:middle line:84% We've taken the object of "devour" 00:46:04.910 --> 00:46:08.400 align:middle line:84% and moved it in the spec of TP, because TP needs to specify. 00:46:08.400 --> 00:46:08.900 align:middle line:90% Yeah? 00:46:08.900 --> 00:46:11.600 align:middle line:90% 00:46:11.600 --> 00:46:13.280 align:middle line:84% We've seen that not all languages have 00:46:13.280 --> 00:46:14.750 align:middle line:90% all of these kinds of movement. 00:46:14.750 --> 00:46:17.420 align:middle line:84% So there are languages with wh-in-situ. 00:46:17.420 --> 00:46:19.625 align:middle line:90% This is the Chaha example. 00:46:19.625 --> 00:46:21.630 align:middle line:84% Chaha is a language spoken in Ethiopia. 00:46:21.630 --> 00:46:23.480 align:middle line:90% It's a Semitic language. 00:46:23.480 --> 00:46:28.310 align:middle line:84% It was actually-- I teach the graduate field methods 00:46:28.310 --> 00:46:36.067 align:middle line:84% class and Chaha was the first language that we worked on. 00:46:36.067 --> 00:46:38.150 align:middle line:84% That's the one-- I think I've told you this story. 00:46:38.150 --> 00:46:41.600 align:middle line:84% That's the one in which I tried to start the class by saying 00:46:41.600 --> 00:46:45.500 align:middle line:84% "The man cooked the meat," and the guy who we were working 00:46:45.500 --> 00:46:47.100 align:middle line:84% with said, "No, I cannot say that." 00:46:47.100 --> 00:46:47.880 align:middle line:90% And I said, "Why?" 00:46:47.880 --> 00:46:49.370 align:middle line:90% He said, "Men do not cook." 00:46:49.370 --> 00:46:51.170 align:middle line:90% It was like, oh, OK. 00:46:51.170 --> 00:46:55.470 align:middle line:84% So, Chaha-- he also taught us an expression for "very quickly," 00:46:55.470 --> 00:46:58.883 align:middle line:84% which meant "in the time it takes spit to evaporate," 00:46:58.883 --> 00:47:00.800 align:middle line:84% which I thought was a cool expression to have. 00:47:00.800 --> 00:47:02.842 align:middle line:84% I wish I could remember how to say that in Chaha. 00:47:02.842 --> 00:47:04.560 align:middle line:90% I should go look it up. 00:47:04.560 --> 00:47:06.440 align:middle line:84% So Chahah is a wh-in-situ language. 00:47:06.440 --> 00:47:08.970 align:middle line:84% It leaves the wh-phrase where objects normally go. 00:47:08.970 --> 00:47:10.340 align:middle line:90% It's also a head final language. 00:47:10.340 --> 00:47:14.660 align:middle line:84% It has the same word order as Japanese. 00:47:14.660 --> 00:47:16.170 align:middle line:90% It's a wh-in-situ language. 00:47:16.170 --> 00:47:18.140 align:middle line:84% So there are languages that leave wh in-situ. 00:47:18.140 --> 00:47:19.580 align:middle line:90% They don't move it. 00:47:19.580 --> 00:47:22.130 align:middle line:84% There are languages that don't go in for NP movement, 00:47:22.130 --> 00:47:23.630 align:middle line:84% at least as much as English does. 00:47:23.630 --> 00:47:27.440 align:middle line:84% So in English, I would have to say, 00:47:27.440 --> 00:47:29.090 align:middle line:90% "The cookies have been eaten." 00:47:29.090 --> 00:47:31.760 align:middle line:84% But in Italian, you can say something that literally means 00:47:31.760 --> 00:47:33.200 align:middle line:90% "Have been eaten the cookies." 00:47:33.200 --> 00:47:35.480 align:middle line:84% So the cookies can just stay in object position. 00:47:35.480 --> 00:47:37.550 align:middle line:90% The Italians don't have the EPP. 00:47:37.550 --> 00:47:40.460 align:middle line:84% They're very relaxed about whether TP has a specifier 00:47:40.460 --> 00:47:42.260 align:middle line:84% or not, so the cookies can just stay where. 00:47:42.260 --> 00:47:44.400 align:middle line:84% They are they don't have to go anywhere. 00:47:44.400 --> 00:47:46.550 align:middle line:84% So Chaha doesn't have to move its wh-phrases. 00:47:46.550 --> 00:47:48.717 align:middle line:84% Italian doesn't have to move its NPs. 00:47:48.717 --> 00:47:51.050 align:middle line:84% There are also cross-linguistic differences with respect 00:47:51.050 --> 00:47:52.980 align:middle line:90% to where heads go. 00:47:52.980 --> 00:47:58.340 align:middle line:84% So here's an English sentence, "Mary often speaks French," 00:47:58.340 --> 00:48:00.680 align:middle line:90% where I've put an adverb. 00:48:00.680 --> 00:48:03.050 align:middle line:90% I've adjoined-- it's an adjunct. 00:48:03.050 --> 00:48:06.950 align:middle line:84% It's not selected by anything, that adverb phrase, "often." 00:48:06.950 --> 00:48:09.240 align:middle line:84% I've attached it to the verb phrase there. 00:48:09.240 --> 00:48:12.920 align:middle line:84% And then you've got a verb with its object, "speaks French." 00:48:12.920 --> 00:48:16.880 align:middle line:84% If any of you speak French-- does anybody speak French? 00:48:16.880 --> 00:48:19.430 align:middle line:84% So if any of you know any French-- a couple of you do, 00:48:19.430 --> 00:48:20.090 align:middle line:90% it looks like. 00:48:20.090 --> 00:48:20.998 align:middle line:90% Maybe you know. 00:48:20.998 --> 00:48:22.790 align:middle line:84% This isn't the right word order for French. 00:48:22.790 --> 00:48:25.430 align:middle line:84% The right word order for French involves 00:48:25.430 --> 00:48:28.790 align:middle line:84% taking the word for "speaks" and putting it before "often." 00:48:28.790 --> 00:48:32.780 align:middle line:84% You can't, in French, say "Mary often speaks French." 00:48:32.780 --> 00:48:36.830 align:middle line:84% You must literally say, "Mary speaks often French," 00:48:36.830 --> 00:48:39.050 align:middle line:90% which in English, we cannot say. 00:48:39.050 --> 00:48:43.850 align:middle line:84% So this is another basic difference between languages. 00:48:43.850 --> 00:48:46.790 align:middle line:84% French requires the verb to raise to T. Sorry, 00:48:46.790 --> 00:48:47.720 align:middle line:90% I called it I again. 00:48:47.720 --> 00:48:49.850 align:middle line:90% I'll try to fix that. 00:48:49.850 --> 00:48:51.500 align:middle line:84% French requires the verb to raise 00:48:51.500 --> 00:48:54.140 align:middle line:84% to T when there's no auxiliary, so you 00:48:54.140 --> 00:48:57.200 align:middle line:90% say "Mary speaks often French." 00:48:57.200 --> 00:48:59.690 align:middle line:84% English doesn't require the verb to raise to T. 00:48:59.690 --> 00:49:04.880 align:middle line:84% So just like English requires the wh-phrase to move 00:49:04.880 --> 00:49:07.190 align:middle line:84% to the specifier of CP and Chaha doesn't-- 00:49:07.190 --> 00:49:09.950 align:middle line:84% Mandarin doesn't, Japanese doesn't-- 00:49:09.950 --> 00:49:14.600 align:middle line:84% French requires the verb to raise to T, to move to T. 00:49:14.600 --> 00:49:15.432 align:middle line:90% English doesn't. 00:49:15.432 --> 00:49:17.390 align:middle line:84% Yes, so there are differences between languages 00:49:17.390 --> 00:49:20.060 align:middle line:84% along these lines, where one language will have 00:49:20.060 --> 00:49:21.780 align:middle line:84% one set of movements and another language 00:49:21.780 --> 00:49:23.238 align:middle line:84% will have another set of movements. 00:49:23.238 --> 00:49:24.500 align:middle line:90% That's a thing we find. 00:49:24.500 --> 00:49:27.060 align:middle line:84% It would be nice to know why and try to derive this 00:49:27.060 --> 00:49:31.020 align:middle line:84% from something else, but that's where the field is right now. 00:49:31.020 --> 00:49:33.480 align:middle line:84% I haven't talked very much about why 00:49:33.480 --> 00:49:36.847 align:middle line:84% these movements happen, and I won't, mainly 00:49:36.847 --> 00:49:37.680 align:middle line:90% for reasons of time. 00:49:37.680 --> 00:49:38.860 align:middle line:90% It's a very interesting topic. 00:49:38.860 --> 00:49:39.870 align:middle line:84% There's a lot of work on it in trying 00:49:39.870 --> 00:49:42.078 align:middle line:84% to figure out what it is that drives these movements, 00:49:42.078 --> 00:49:43.210 align:middle line:90% causes them to happen. 00:49:43.210 --> 00:49:44.960 align:middle line:84% So the only one that I've really given you 00:49:44.960 --> 00:49:48.630 align:middle line:84% a motivation for is NP-movement, which I've said 00:49:48.630 --> 00:49:51.572 align:middle line:84% is driven by the EPP, the need for TP 00:49:51.572 --> 00:49:53.280 align:middle line:84% to have a specifier, which is a need that 00:49:53.280 --> 00:49:55.280 align:middle line:84% exists in some languages, but not others. 00:49:55.280 --> 00:49:59.040 align:middle line:84% So English has that, and Italian, for example, doesn't. 00:49:59.040 --> 00:50:01.620 align:middle line:84% But I do want to talk some-- so I haven't talked about why 00:50:01.620 --> 00:50:02.310 align:middle line:90% these movements happen. 00:50:02.310 --> 00:50:04.477 align:middle line:84% But I want to talk some about some of the conditions 00:50:04.477 --> 00:50:05.040 align:middle line:90% on movement. 00:50:05.040 --> 00:50:07.110 align:middle line:84% Because what we're going to see is that movement can't always 00:50:07.110 --> 00:50:09.585 align:middle line:84% happen, and we're going to want to try to understand why. 00:50:09.585 --> 00:50:13.970 align:middle line:90% 00:50:13.970 --> 00:50:16.870 align:middle line:84% So you can say things like "I ordered 00:50:16.870 --> 00:50:21.180 align:middle line:84% a hamburger and French fries," but it would be weird 00:50:21.180 --> 00:50:23.400 align:middle line:84% for me to ask you a question like, "What did you 00:50:23.400 --> 00:50:25.880 align:middle line:90% order a hamburger and?" 00:50:25.880 --> 00:50:27.140 align:middle line:90% Do people agree? 00:50:27.140 --> 00:50:29.418 align:middle line:84% That's a strange question to ask. 00:50:29.418 --> 00:50:31.460 align:middle line:84% So it's sometimes called the coordinate structure 00:50:31.460 --> 00:50:33.000 align:middle line:90% constraint. 00:50:33.000 --> 00:50:35.150 align:middle line:84% And it's one of many examples where 00:50:35.150 --> 00:50:41.390 align:middle line:84% people have found the wh-movement example is somehow 00:50:41.390 --> 00:50:42.590 align:middle line:90% blocked. 00:50:42.590 --> 00:50:45.740 align:middle line:84% So all of the wh-movement examples I've shown you so far, 00:50:45.740 --> 00:50:48.410 align:middle line:84% it's always been possible to move the wh-phrase. 00:50:48.410 --> 00:50:50.100 align:middle line:84% There are times when it's not possible, 00:50:50.100 --> 00:50:51.270 align:middle line:90% and this is one example. 00:50:51.270 --> 00:50:53.720 align:middle line:84% So "What did you order a hamburger and?" 00:50:53.720 --> 00:50:54.350 align:middle line:90% No good. 00:50:54.350 --> 00:50:57.470 align:middle line:84% It's roughly if you have two things that 00:50:57.470 --> 00:51:00.650 align:middle line:84% are connected by "and," you can't wh-move one of them. 00:51:00.650 --> 00:51:04.407 align:middle line:84% You also can't say "What did you order and a hamburger," 00:51:04.407 --> 00:51:06.740 align:middle line:84% where you moved the first one instead of the second one. 00:51:06.740 --> 00:51:07.670 align:middle line:90% Also no good. 00:51:07.670 --> 00:51:11.200 align:middle line:90% 00:51:11.200 --> 00:51:13.030 align:middle line:90% Yeah? 00:51:13.030 --> 00:51:14.540 align:middle line:90% Yeah? 00:51:14.540 --> 00:51:16.540 align:middle line:84% AUDIENCE: The second one sounded fine, actually. 00:51:16.540 --> 00:51:18.248 align:middle line:84% NORVIN RICHARDS: I'm sorry, say it again? 00:51:18.248 --> 00:51:19.992 align:middle line:90% The second one sounded fine? 00:51:19.992 --> 00:51:20.950 align:middle line:90% AUDIENCE: I don't know. 00:51:20.950 --> 00:51:22.742 align:middle line:84% Like, "What did you order a hamburger and?" 00:51:22.742 --> 00:51:25.360 align:middle line:90% 00:51:25.360 --> 00:51:27.580 align:middle line:84% Like, I have to somehow think, oh, 00:51:27.580 --> 00:51:29.543 align:middle line:84% that is not grammatically correct. 00:51:29.543 --> 00:51:30.460 align:middle line:90% NORVIN RICHARDS: Yeah? 00:51:30.460 --> 00:51:34.462 align:middle line:84% AUDIENCE: It's just on-- like, if someone just asked me 00:51:34.462 --> 00:51:36.877 align:middle line:90% that, I would just [INAUDIBLE]. 00:51:36.877 --> 00:51:41.230 align:middle line:90% 00:51:41.230 --> 00:51:43.060 align:middle line:90% It's somehow-- no, the first on. 00:51:43.060 --> 00:51:43.420 align:middle line:90% NORVIN RICHARDS: The first one. 00:51:43.420 --> 00:51:44.080 align:middle line:90% Yeah, this one. 00:51:44.080 --> 00:51:45.430 align:middle line:84% "What did you order a hamburger and?" 00:51:45.430 --> 00:51:47.388 align:middle line:84% You mean if somebody asked you that, you'd just 00:51:47.388 --> 00:51:48.537 align:middle line:90% be like, "French fries." 00:51:48.537 --> 00:51:49.120 align:middle line:90% AUDIENCE: Huh? 00:51:49.120 --> 00:51:51.078 align:middle line:84% NORVIN RICHARDS: You'd just say "French fries." 00:51:51.078 --> 00:51:52.560 align:middle line:90% AUDIENCE: Yeah. 00:51:52.560 --> 00:51:54.720 align:middle line:84% Like, now that I'm listening to it, 00:51:54.720 --> 00:51:56.170 align:middle line:90% it's grammatically incorrect. 00:51:56.170 --> 00:51:57.420 align:middle line:90% Now that I'm playing it back. 00:51:57.420 --> 00:51:57.870 align:middle line:90% NORVIN RICHARDS: Yeah. 00:51:57.870 --> 00:52:00.105 align:middle line:84% AUDIENCE: But it just somehow doesn't sound weird. 00:52:00.105 --> 00:52:00.600 align:middle line:90% NORVIN RICHARDS: Hm. 00:52:00.600 --> 00:52:01.600 align:middle line:90% Oh, that's interesting. 00:52:01.600 --> 00:52:04.265 align:middle line:84% I mean, several other people are raising their hand, 00:52:04.265 --> 00:52:06.390 align:middle line:84% so maybe I'll just open this to general discussion. 00:52:06.390 --> 00:52:07.925 align:middle line:90% Joseph? 00:52:07.925 --> 00:52:09.300 align:middle line:84% AUDIENCE: I feel like it would be 00:52:09.300 --> 00:52:11.112 align:middle line:84% natural to put a colon after "order." 00:52:11.112 --> 00:52:13.320 align:middle line:84% I don't know what that would do, like, syntactically, 00:52:13.320 --> 00:52:14.970 align:middle line:90% but "What did you order-- 00:52:14.970 --> 00:52:17.710 align:middle line:90% a hamburger and?" 00:52:17.710 --> 00:52:20.920 align:middle line:84% NORVIN RICHARDS: Yeah, so forget about "and" for a second. 00:52:20.920 --> 00:52:25.220 align:middle line:84% I think it's possible to say, OK, "you ordered--" 00:52:25.220 --> 00:52:26.960 align:middle line:84% and have that be a question, right? 00:52:26.960 --> 00:52:29.520 align:middle line:84% So it's your job to fill in the blank for me. 00:52:29.520 --> 00:52:32.843 align:middle line:84% And I think what you just did is a version of that where 00:52:32.843 --> 00:52:34.010 align:middle line:90% I say, "What did you order?" 00:52:34.010 --> 00:52:35.330 align:middle line:90% That's the wh-question. 00:52:35.330 --> 00:52:37.160 align:middle line:84% And now I'm going to ask you to help 00:52:37.160 --> 00:52:38.450 align:middle line:90% me finish the sentence for me. 00:52:38.450 --> 00:52:39.950 align:middle line:90% I think that's what's going on. 00:52:39.950 --> 00:52:41.090 align:middle line:90% Kateryna? 00:52:41.090 --> 00:52:44.210 align:middle line:84% AUDIENCE: When you say you ordered dot, dot, dot, 00:52:44.210 --> 00:52:50.700 align:middle line:84% is that equivalent to not doing wh-movement and just not 00:52:50.700 --> 00:52:52.840 align:middle line:90% pronouncing the "what?" 00:52:52.840 --> 00:52:56.240 align:middle line:84% NORVIN RICHARDS: It's an interesting question. 00:52:56.240 --> 00:52:59.080 align:middle line:84% I don't know whether we want the syntax to handle that 00:52:59.080 --> 00:53:04.280 align:middle line:84% or not, or whether we want that to be like the kind of case 00:53:04.280 --> 00:53:05.270 align:middle line:90% I talked about before. 00:53:05.270 --> 00:53:09.070 align:middle line:90% 00:53:09.070 --> 00:53:12.410 align:middle line:90% So I-- because I think-- 00:53:12.410 --> 00:53:16.040 align:middle line:84% there's a restriction on that technique. 00:53:16.040 --> 00:53:18.800 align:middle line:84% For example, it has to be the end of the sentence. 00:53:18.800 --> 00:53:21.080 align:middle line:84% You have to stop talking before you get to the end. 00:53:21.080 --> 00:53:26.120 align:middle line:84% I can't say "you ordered and a hamburger," 00:53:26.120 --> 00:53:28.940 align:middle line:90% I don't think, right? 00:53:28.940 --> 00:53:33.890 align:middle line:84% So I think what I'm doing is simulating 00:53:33.890 --> 00:53:37.250 align:middle line:84% forgetting how my sentence was going to end, right? 00:53:37.250 --> 00:53:41.840 align:middle line:84% So it's really as though I'm about to say "You ordered 00:53:41.840 --> 00:53:44.360 align:middle line:84% a hamburger" and I'm pretending that I can't 00:53:44.360 --> 00:53:45.830 align:middle line:84% remember how to end the sentence, 00:53:45.830 --> 00:53:48.780 align:middle line:90% and I'm asking you to help me. 00:53:48.780 --> 00:53:50.510 align:middle line:84% I have a feeling that's the kind of way 00:53:50.510 --> 00:53:54.960 align:middle line:84% we want to think about that kind of example, but yeah. 00:53:54.960 --> 00:53:56.790 align:middle line:90% Raequel? 00:53:56.790 --> 00:53:58.767 align:middle line:84% AUDIENCE: When I hear, "What did you order?" 00:53:58.767 --> 00:54:01.337 align:middle line:84% at least as well, but if someone said, "What did you 00:54:01.337 --> 00:54:02.940 align:middle line:90% order a hamburger with?" 00:54:02.940 --> 00:54:04.190 align:middle line:90% I would be, like, that's fine. 00:54:04.190 --> 00:54:05.357 align:middle line:90% NORVIN RICHARDS: Yeah, yeah. 00:54:05.357 --> 00:54:06.437 align:middle line:90% Yeah, so good point. 00:54:06.437 --> 00:54:08.770 align:middle line:84% And maybe that's got something to do with your reaction, 00:54:08.770 --> 00:54:13.060 align:middle line:84% that you are being maximally charitable to the person you're 00:54:13.060 --> 00:54:13.670 align:middle line:90% talking to. 00:54:13.670 --> 00:54:16.128 align:middle line:84% If somebody said, "What are you ordering a hamburger-- what 00:54:16.128 --> 00:54:18.070 align:middle line:90% did you order a hamburger and?" 00:54:18.070 --> 00:54:20.860 align:middle line:84% you do your best to pretend that they 00:54:20.860 --> 00:54:22.300 align:middle line:90% said something grammatical. 00:54:22.300 --> 00:54:23.733 align:middle line:90% It's clear what they meant. 00:54:23.733 --> 00:54:25.900 align:middle line:84% And so you're going to pretend that they said "with" 00:54:25.900 --> 00:54:27.858 align:middle line:90% instead of "and." 00:54:27.858 --> 00:54:29.650 align:middle line:84% Maybe there's something like that going on. 00:54:29.650 --> 00:54:33.580 align:middle line:84% This is reminding me of a-- there's work on-- 00:54:33.580 --> 00:54:35.180 align:middle line:90% what's it called? 00:54:35.180 --> 00:54:39.100 align:middle line:84% I think it's called the Moses illusion, where people have 00:54:39.100 --> 00:54:43.090 align:middle line:84% observed that if you ask people how many animals of each kind 00:54:43.090 --> 00:54:47.020 align:middle line:84% did Moses take on the Ark, where if you're 00:54:47.020 --> 00:54:49.540 align:middle line:84% familiar with the Bible, you know that Moses did not 00:54:49.540 --> 00:54:51.430 align:middle line:90% take any animals on the Ark. 00:54:51.430 --> 00:54:54.340 align:middle line:84% The Ark was Noah, a different guy. 00:54:54.340 --> 00:54:56.350 align:middle line:84% Noah was the one who took animals on the Ark. 00:54:56.350 --> 00:54:59.470 align:middle line:84% And so the real response to that is supposed to be, hold on, 00:54:59.470 --> 00:55:01.253 align:middle line:84% Moses didn't take any animals on the Ark. 00:55:01.253 --> 00:55:03.670 align:middle line:84% But that if you do a survey in which you walk up to people 00:55:03.670 --> 00:55:05.628 align:middle line:84% at random and say how many animals of each type 00:55:05.628 --> 00:55:08.020 align:middle line:84% did Moses take on the Ark, what they typically do, 00:55:08.020 --> 00:55:10.750 align:middle line:84% very often, is pretend that you said Noah. 00:55:10.750 --> 00:55:13.432 align:middle line:84% So they pretend that you asked the question 00:55:13.432 --> 00:55:14.890 align:middle line:84% in a way that makes sense, and they 00:55:14.890 --> 00:55:16.390 align:middle line:90% tell you-- they tell you two. 00:55:16.390 --> 00:55:19.495 align:middle line:90% Which is actually not-- 00:55:19.495 --> 00:55:21.370 align:middle line:84% as I said, I've been reading the Bible lately 00:55:21.370 --> 00:55:22.787 align:middle line:84% because of this Wampanoag project. 00:55:22.787 --> 00:55:25.270 align:middle line:84% Noah-- I hadn't remembered this, but Noah 00:55:25.270 --> 00:55:26.890 align:middle line:84% took two of some kinds of animals, 00:55:26.890 --> 00:55:28.210 align:middle line:90% but more of other kinds. 00:55:28.210 --> 00:55:31.450 align:middle line:84% It had to do with whether they were edible or not, i think. 00:55:31.450 --> 00:55:35.410 align:middle line:84% So whether the laws for what you're allowed to eat 00:55:35.410 --> 00:55:36.850 align:middle line:90% allowed you to eat them. 00:55:36.850 --> 00:55:38.560 align:middle line:84% If it was OK to eat them, he took more, 00:55:38.560 --> 00:55:41.770 align:middle line:84% I guess so that there would be some to spare. 00:55:41.770 --> 00:55:42.447 align:middle line:90% Yes. 00:55:42.447 --> 00:55:44.530 align:middle line:84% So I suspect that that's what's going on with you. 00:55:44.530 --> 00:55:45.905 align:middle line:84% But it's an interesting question. 00:55:45.905 --> 00:55:47.990 align:middle line:84% Actually, your point raises an interesting point, 00:55:47.990 --> 00:55:50.080 align:middle line:90% which is that these-- 00:55:50.080 --> 00:55:52.570 align:middle line:84% we're going to talk about several cases of this kind 00:55:52.570 --> 00:55:57.250 align:middle line:84% where a wh-phrase seems to be constrained from moving 00:55:57.250 --> 00:55:59.710 align:middle line:84% from a particular place, and the examples 00:55:59.710 --> 00:56:01.810 align:middle line:84% are always going to be examples, hopefully, 00:56:01.810 --> 00:56:04.060 align:middle line:84% where it's not that you can't figure out 00:56:04.060 --> 00:56:05.380 align:middle line:90% what the other person means. 00:56:05.380 --> 00:56:07.030 align:middle line:84% It's not that these are unthinkable. 00:56:07.030 --> 00:56:09.490 align:middle line:90% They're just unsayable. 00:56:09.490 --> 00:56:11.320 align:middle line:84% It's not the right way to say it. 00:56:11.320 --> 00:56:12.820 align:middle line:84% Actually, the next example, I think, 00:56:12.820 --> 00:56:15.560 align:middle line:84% is particularly clear in that regard. 00:56:15.560 --> 00:56:16.960 align:middle line:90% So here are two sentences. 00:56:16.960 --> 00:56:18.943 align:middle line:84% We've talked a lot about embedded clauses. 00:56:18.943 --> 00:56:20.860 align:middle line:84% And I guess when we were talking about German, 00:56:20.860 --> 00:56:24.040 align:middle line:84% I touched on the fact that embedded clauses sometimes have 00:56:24.040 --> 00:56:26.440 align:middle line:90% it start with complementizers. 00:56:26.440 --> 00:56:28.480 align:middle line:84% Our English sentences have pretty reliably 00:56:28.480 --> 00:56:31.240 align:middle line:84% started with complementizers, but they don't have to, right? 00:56:31.240 --> 00:56:32.868 align:middle line:84% So you can say, "I think that Mary 00:56:32.868 --> 00:56:34.910 align:middle line:84% should win the election," where there's a "that." 00:56:34.910 --> 00:56:37.452 align:middle line:84% But you can also say, "I think Mary should win the election." 00:56:37.452 --> 00:56:39.310 align:middle line:90% There's no complementizer there. 00:56:39.310 --> 00:56:43.600 align:middle line:84% And then linguists get very curious about whether, OK-- 00:56:43.600 --> 00:56:46.300 align:middle line:84% I got asked this question in class at one point. 00:56:46.300 --> 00:56:49.330 align:middle line:84% Does that mean that I should draw different trees for these? 00:56:49.330 --> 00:56:51.580 align:middle line:84% Or does it mean that the embedded clause should 00:56:51.580 --> 00:56:56.050 align:middle line:84% be a CP in both cases, and that C can either be pronounced 00:56:56.050 --> 00:56:58.680 align:middle line:90% "that" or [SMACKS LIPS]. 00:56:58.680 --> 00:56:59.190 align:middle line:90% Right? 00:56:59.190 --> 00:57:03.450 align:middle line:84% That C has a pronunciation where it's just not pronounced. 00:57:03.450 --> 00:57:06.548 align:middle line:84% And that's a debate that people have. 00:57:06.548 --> 00:57:07.590 align:middle line:90% We don't have to have it. 00:57:07.590 --> 00:57:09.480 align:middle line:84% But let me call your attention to a fact which 00:57:09.480 --> 00:57:11.438 align:middle line:84% doesn't hold actually for all English speakers, 00:57:11.438 --> 00:57:14.125 align:middle line:90% but for some it does. 00:57:14.125 --> 00:57:16.500 align:middle line:84% In my English, it's OK to ask questions like, "Who do you 00:57:16.500 --> 00:57:18.670 align:middle line:90% think should win the election?" 00:57:18.670 --> 00:57:20.860 align:middle line:84% But questions like, "Who do you think 00:57:20.860 --> 00:57:24.164 align:middle line:90% that should win the election?" 00:57:24.164 --> 00:57:26.740 align:middle line:90% are no good for me. 00:57:26.740 --> 00:57:29.178 align:middle line:84% Am I the only person like that here? 00:57:29.178 --> 00:57:31.095 align:middle line:84% Is there anybody for whom these are both fine? 00:57:31.095 --> 00:57:34.700 align:middle line:90% 00:57:34.700 --> 00:57:36.860 align:middle line:84% Are you communicating with me in ASL or-- 00:57:36.860 --> 00:57:38.810 align:middle line:84% AUDIENCE: I do not know what the aslerisk-- 00:57:38.810 --> 00:57:40.220 align:middle line:84% NORVIN RICHARDS: The asterisk means it's bad. 00:57:40.220 --> 00:57:41.200 align:middle line:84% AUDIENCE: The first one feels fine. 00:57:41.200 --> 00:57:42.670 align:middle line:90% The second one feels not fine. 00:57:42.670 --> 00:57:43.753 align:middle line:90% NORVIN RICHARDS: OK, good. 00:57:43.753 --> 00:57:45.280 align:middle line:90% Yeah, so you're like me. 00:57:45.280 --> 00:57:46.750 align:middle line:90% I mean, interesting, yeah. 00:57:46.750 --> 00:57:49.480 align:middle line:84% It would be OK if that were not true, but yeah. 00:57:49.480 --> 00:57:52.030 align:middle line:84% There are people for whom both of these are fine. 00:57:52.030 --> 00:57:54.160 align:middle line:84% So occasionally, when I'm teaching this class, 00:57:54.160 --> 00:57:56.710 align:middle line:84% I'll do this slide, and I'll look out in the audience 00:57:56.710 --> 00:57:59.080 align:middle line:84% and there'll be one person who looks really alarmed, 00:57:59.080 --> 00:58:00.627 align:middle line:90% like they're hallucinating. 00:58:00.627 --> 00:58:02.710 align:middle line:84% And I'm like, you're from the Midwest, aren't you? 00:58:02.710 --> 00:58:04.390 align:middle line:90% And they're like, yes. 00:58:04.390 --> 00:58:07.910 align:middle line:84% So there's a part of America where they do this. 00:58:07.910 --> 00:58:10.383 align:middle line:84% It's something about living in big, flat areas 00:58:10.383 --> 00:58:11.800 align:middle line:84% with lots of agriculture, I guess, 00:58:11.800 --> 00:58:15.530 align:middle line:84% that it's OK to say these things. 00:58:15.530 --> 00:58:17.795 align:middle line:84% There's a lot of interesting questions about why, 00:58:17.795 --> 00:58:20.170 align:middle line:84% or what it is that distinguishes some dialects of English 00:58:20.170 --> 00:58:20.830 align:middle line:90% from others. 00:58:20.830 --> 00:58:24.520 align:middle line:84% Because it's not as if you cover this 00:58:24.520 --> 00:58:26.270 align:middle line:84% in high school, or your mom and dad 00:58:26.270 --> 00:58:28.270 align:middle line:84% punish if you say, "Who do you think that should 00:58:28.270 --> 00:58:29.140 align:middle line:90% win the election," right? 00:58:29.140 --> 00:58:30.140 align:middle line:90% That's not what happens. 00:58:30.140 --> 00:58:34.720 align:middle line:84% But for most Native speakers of English, there's this contrast. 00:58:34.720 --> 00:58:37.300 align:middle line:84% It's called the that-trace effect. 00:58:37.300 --> 00:58:38.200 align:middle line:90% Never mind why. 00:58:38.200 --> 00:58:42.340 align:middle line:84% Well, because the idea is it's impossible to have 00:58:42.340 --> 00:58:45.550 align:middle line:84% a "that" immediately followed by the place 00:58:45.550 --> 00:58:46.780 align:middle line:90% the wh-phrase came from. 00:58:46.780 --> 00:58:49.150 align:middle line:84% If you imagine the wh-phrase when it moves, 00:58:49.150 --> 00:58:51.650 align:middle line:84% it leaves traces of itself behind, I guess, 00:58:51.650 --> 00:58:52.460 align:middle line:90% is the metaphor. 00:58:52.460 --> 00:58:52.960 align:middle line:90% Yeah. 00:58:52.960 --> 00:58:56.620 align:middle line:90% 00:58:56.620 --> 00:58:59.242 align:middle line:90% This is specifically about-- 00:58:59.242 --> 00:59:01.450 align:middle line:84% I guess this came out in the way I just described it. 00:59:01.450 --> 00:59:04.100 align:middle line:84% It's specifically about extraction of subjects. 00:59:04.100 --> 00:59:05.650 align:middle line:84% So it's bad to say, "Who do you think 00:59:05.650 --> 00:59:07.188 align:middle line:90% that should win the election?" 00:59:07.188 --> 00:59:08.980 align:middle line:84% But it's OK to say things like, "Who do you 00:59:08.980 --> 00:59:10.272 align:middle line:90% think that we should vote for?" 00:59:10.272 --> 00:59:14.350 align:middle line:84% So the generalization is not you can't wh-move past a "that." 00:59:14.350 --> 00:59:16.210 align:middle line:84% it's OK to wh-move past a "that." 00:59:16.210 --> 00:59:18.050 align:middle line:84% It's just not OK to wh-move past "that" 00:59:18.050 --> 00:59:20.440 align:middle line:90% which is right next to you. 00:59:20.440 --> 00:59:23.650 align:middle line:84% And lots of interesting work on what the heck is going on, 00:59:23.650 --> 00:59:24.700 align:middle line:90% yeah. 00:59:24.700 --> 00:59:26.312 align:middle line:84% Again, let me just urge you, if you 00:59:26.312 --> 00:59:27.770 align:middle line:84% would like to know more about this, 00:59:27.770 --> 00:59:29.080 align:middle line:84% go take more linguistics classes. 00:59:29.080 --> 00:59:30.400 align:middle line:84% This is something linguists work on, 00:59:30.400 --> 00:59:31.942 align:middle line:84% trying to figure out what's going on. 00:59:31.942 --> 00:59:36.070 align:middle line:90% 00:59:36.070 --> 00:59:39.400 align:middle line:84% Another example of movement not being possible. 00:59:39.400 --> 00:59:41.530 align:middle line:84% We actually had an example like this earlier. 00:59:41.530 --> 00:59:44.260 align:middle line:84% All of our CPs so far, or most of our CPs 00:59:44.260 --> 00:59:46.102 align:middle line:84% so far have been complements of verbs. 00:59:46.102 --> 00:59:47.560 align:middle line:84% So we've said things like, "I think 00:59:47.560 --> 00:59:49.353 align:middle line:84% that she should win the election," where 00:59:49.353 --> 00:59:50.770 align:middle line:84% "that she should win the election" 00:59:50.770 --> 00:59:52.210 align:middle line:90% is the complement of "think." 00:59:52.210 --> 00:59:55.330 align:middle line:84% Or "people believe that the moon is made of green cheese," 00:59:55.330 --> 00:59:59.380 align:middle line:84% where "believe" takes as its sister, its complement, a CP, 00:59:59.380 --> 01:00:01.540 align:middle line:84% "that the moon is made of green cheese." 01:00:01.540 --> 01:00:03.880 align:middle line:84% It's also possible for CPs to be subjects, 01:00:03.880 --> 01:00:05.980 align:middle line:84% as in "That the moon is made of green cheese 01:00:05.980 --> 01:00:07.120 align:middle line:90% is widely believed." 01:00:07.120 --> 01:00:10.210 align:middle line:90% That's false, but grammatical. 01:00:10.210 --> 01:00:11.810 align:middle line:90% People agree? 01:00:11.810 --> 01:00:12.310 align:middle line:90% OK. 01:00:12.310 --> 01:00:15.160 align:middle line:84% It's not the most natural thing I've said today, 01:00:15.160 --> 01:00:18.320 align:middle line:90% but it's grammatical. 01:00:18.320 --> 01:00:23.350 align:middle line:84% Now, so there's a tree for "That the moon is 01:00:23.350 --> 01:00:25.360 align:middle line:84% made of green cheese is widely believed." 01:00:25.360 --> 01:00:28.090 align:middle line:84% We've got a CP, which is sitting in the specifier of a TP. 01:00:28.090 --> 01:00:30.850 align:middle line:84% It's the subject of "is widely believed." 01:00:30.850 --> 01:00:35.470 align:middle line:90% 01:00:35.470 --> 01:00:36.940 align:middle line:90% Now let's do some wh-movement. 01:00:36.940 --> 01:00:39.190 align:middle line:84% It's OK to ask questions like, "What do people believe 01:00:39.190 --> 01:00:40.870 align:middle line:90% that the moon is made of?" 01:00:40.870 --> 01:00:44.140 align:middle line:84% so we took that first sentence and we wh-moved. 01:00:44.140 --> 01:00:46.557 align:middle line:84% We turned "green cheese" into "what" and wh-moved it. 01:00:46.557 --> 01:00:48.640 align:middle line:84% "What do people believe that the moon is made of?" 01:00:48.640 --> 01:00:49.890 align:middle line:90% Fine. 01:00:49.890 --> 01:00:52.430 align:middle line:84% But you can see where this is going. 01:00:52.430 --> 01:00:57.770 align:middle line:84% "What is that the moon is made of widely believed?" 01:00:57.770 --> 01:00:59.720 align:middle line:90% Ow, yeah? 01:00:59.720 --> 01:01:01.130 align:middle line:90% So, you know. 01:01:01.130 --> 01:01:03.980 align:middle line:84% "That the moon is made of green cheese is widely believed," OK, 01:01:03.980 --> 01:01:05.780 align:middle line:84% that's possibly a little awkward. 01:01:05.780 --> 01:01:09.720 align:middle line:84% Not something I'm likely to say in casual conversation. 01:01:09.720 --> 01:01:12.770 align:middle line:84% But the last example is just word salad, 01:01:12.770 --> 01:01:13.970 align:middle line:90% to use the technical term. 01:01:13.970 --> 01:01:14.720 align:middle line:90% Yeah. 01:01:14.720 --> 01:01:18.200 align:middle line:84% It's very unclear what the heck it's supposed to mean. 01:01:18.200 --> 01:01:19.550 align:middle line:90% Do people agree? 01:01:19.550 --> 01:01:22.880 align:middle line:84% Again, I'm reporting my memory of judgments 01:01:22.880 --> 01:01:26.060 align:middle line:84% that I had when I was young and carefree, 01:01:26.060 --> 01:01:27.680 align:middle line:84% but I'm pretty sure this is true. 01:01:27.680 --> 01:01:29.370 align:middle line:90% You just can't say these things. 01:01:29.370 --> 01:01:30.710 align:middle line:90% Yeah? 01:01:30.710 --> 01:01:31.490 align:middle line:90% OK. 01:01:31.490 --> 01:01:34.010 align:middle line:84% All right, so these kinds of examples 01:01:34.010 --> 01:01:38.540 align:middle line:84% are all just meant to show you there are some kinds of things 01:01:38.540 --> 01:01:40.410 align:middle line:90% that you can't wh-move out of. 01:01:40.410 --> 01:01:43.220 align:middle line:84% So if you coordinate two things, you can't move one of them. 01:01:43.220 --> 01:01:45.838 align:middle line:84% The that-trace effect is a condition on extraction 01:01:45.838 --> 01:01:46.880 align:middle line:90% of subjects specifically. 01:01:46.880 --> 01:01:50.850 align:middle line:84% You can't move those if there's a "that" right to their left. 01:01:50.850 --> 01:01:53.030 align:middle line:84% And if you have a clause which is 01:01:53.030 --> 01:01:56.090 align:middle line:84% a subject, like "that the moon is made of green cheese 01:01:56.090 --> 01:01:59.720 align:middle line:84% is widely believed," you can't wh-extract out of that. 01:01:59.720 --> 01:02:01.400 align:middle line:84% There's a metaphor that's often used 01:02:01.400 --> 01:02:02.810 align:middle line:90% for these kinds of restrictions. 01:02:02.810 --> 01:02:06.870 align:middle line:84% The things that you can't move out of are called islands. 01:02:06.870 --> 01:02:10.880 align:middle line:84% So this last one I've identified for you is a subject island. 01:02:10.880 --> 01:02:14.060 align:middle line:84% So I guess the idea is supposed to be that 01:02:14.060 --> 01:02:17.180 align:middle line:84% if you're a wh-phrase and you're on an island, 01:02:17.180 --> 01:02:19.190 align:middle line:90% wh-phrases cannot swim. 01:02:19.190 --> 01:02:22.047 align:middle line:84% There is no boat, there is no bridge. 01:02:22.047 --> 01:02:23.630 align:middle line:84% If you're on an island, you're doomed. 01:02:23.630 --> 01:02:26.000 align:middle line:90% You cannot get out. 01:02:26.000 --> 01:02:27.510 align:middle line:90% You're stuck. 01:02:27.510 --> 01:02:31.400 align:middle line:84% So there's a thriving literature in syntax 01:02:31.400 --> 01:02:33.770 align:middle line:84% in which people identify islands and try 01:02:33.770 --> 01:02:37.145 align:middle line:84% to figure out why the things are islands that are islands. 01:02:37.145 --> 01:02:39.770 align:middle line:84% We hope that this will teach us some things about the mechanics 01:02:39.770 --> 01:02:40.640 align:middle line:90% of extraction. 01:02:40.640 --> 01:02:41.690 align:middle line:90% Yeah. 01:02:41.690 --> 01:02:44.960 align:middle line:84% I want to show you a particular class of island effects, 01:02:44.960 --> 01:02:48.350 align:middle line:84% or cases where movement is impossible, 01:02:48.350 --> 01:02:50.300 align:middle line:84% and show you a little bit of the work 01:02:50.300 --> 01:02:53.150 align:middle line:84% that people have done to try to make some progress on why 01:02:53.150 --> 01:02:56.300 align:middle line:84% these particular things block extraction. 01:02:56.300 --> 01:02:59.360 align:middle line:84% I think we have time to at least get started on this, 01:02:59.360 --> 01:03:03.363 align:middle line:84% and then we might have to finish it next time. 01:03:03.363 --> 01:03:05.780 align:middle line:84% So there are a number of kinds of restrictions on movement 01:03:05.780 --> 01:03:08.780 align:middle line:84% that could be unified, and I'll show these to you in a second. 01:03:08.780 --> 01:03:12.323 align:middle line:84% They could be unified into a single condition, which 01:03:12.323 --> 01:03:13.490 align:middle line:90% we could call shortest move. 01:03:13.490 --> 01:03:16.610 align:middle line:84% It says if you have a choice between two different movement 01:03:16.610 --> 01:03:19.665 align:middle line:84% operations, you should pick the one that's shorter. 01:03:19.665 --> 01:03:22.040 align:middle line:84% And people are kind of excited by that, because it sounds 01:03:22.040 --> 01:03:24.740 align:middle line:90% like cognitively plausible. 01:03:24.740 --> 01:03:27.530 align:middle line:84% If you're trying to decide between movement operations, 01:03:27.530 --> 01:03:29.630 align:middle line:84% you should pick a short one over a long one. 01:03:29.630 --> 01:03:32.340 align:middle line:90% 01:03:32.340 --> 01:03:34.250 align:middle line:84% I'm going to show you a definition of "short" 01:03:34.250 --> 01:03:35.930 align:middle line:90% in a second. 01:03:35.930 --> 01:03:39.950 align:middle line:84% Please don't go tattoo this on your arm or anything. 01:03:39.950 --> 01:03:42.513 align:middle line:84% There's no-- it's not all that important that we define it 01:03:42.513 --> 01:03:44.180 align:middle line:84% carefully for the purposes of what we're 01:03:44.180 --> 01:03:45.740 align:middle line:90% going to do in this class. 01:03:45.740 --> 01:03:47.655 align:middle line:84% In theoretical work on this topic, 01:03:47.655 --> 01:03:49.280 align:middle line:84% it is important to define it carefully, 01:03:49.280 --> 01:03:50.630 align:middle line:84% and there's work on trying to figure out 01:03:50.630 --> 01:03:51.980 align:middle line:90% exactly how to define it. 01:03:51.980 --> 01:03:53.510 align:middle line:90% But here's one definition. 01:03:53.510 --> 01:03:56.570 align:middle line:90% You could say, take the path-- 01:03:56.570 --> 01:03:58.430 align:middle line:84% take a movement operation, and we'll 01:03:58.430 --> 01:04:02.570 align:middle line:84% consider-- let me show you a movement operation. 01:04:02.570 --> 01:04:04.070 align:middle line:84% So here's a movement operation where 01:04:04.070 --> 01:04:06.770 align:middle line:84% we took "with chopsticks" in this V2 example 01:04:06.770 --> 01:04:10.100 align:middle line:84% and we moved it out of the TP and into the C. 01:04:10.100 --> 01:04:12.140 align:middle line:84% And we'll talk about the path of a movement, 01:04:12.140 --> 01:04:15.410 align:middle line:84% and that'll be the set of nodes that dominate 01:04:15.410 --> 01:04:17.180 align:middle line:84% the original position, the position 01:04:17.180 --> 01:04:20.280 align:middle line:84% you moved out of, and don't dominate the landing site. 01:04:20.280 --> 01:04:24.030 align:middle line:84% So in this particular case of movement, 01:04:24.030 --> 01:04:26.062 align:middle line:84% this CP dominates the landing site. 01:04:26.062 --> 01:04:27.770 align:middle line:84% But then there are a bunch of other nodes 01:04:27.770 --> 01:04:29.395 align:middle line:84% like this one and this one, and a bunch 01:04:29.395 --> 01:04:32.990 align:middle line:84% of other ones inside this TP, that dominate this XP here. 01:04:32.990 --> 01:04:36.980 align:middle line:84% And those are what we'll refer to as the path. 01:04:36.980 --> 01:04:39.890 align:middle line:84% And to claim that a move has to be as short as possible 01:04:39.890 --> 01:04:42.110 align:middle line:84% is to claim that you want the path to be, 01:04:42.110 --> 01:04:44.700 align:middle line:90% well, as small as possible. 01:04:44.700 --> 01:04:46.308 align:middle line:84% And there's some interesting work 01:04:46.308 --> 01:04:47.850 align:middle line:84% on trying to figure out what happens. 01:04:47.850 --> 01:04:50.210 align:middle line:84% So in the cases I'm going to show you, 01:04:50.210 --> 01:04:52.670 align:middle line:84% the two paths are in a subset relation. 01:04:52.670 --> 01:04:55.130 align:middle line:84% One of them just contains a subset 01:04:55.130 --> 01:04:57.200 align:middle line:84% of the nodes that are in the other one. 01:04:57.200 --> 01:04:58.160 align:middle line:84% And there's an interesting question 01:04:58.160 --> 01:04:59.827 align:middle line:84% about what happens if they just overlap? 01:04:59.827 --> 01:05:01.100 align:middle line:90% Do you actually count nodes? 01:05:01.100 --> 01:05:02.642 align:middle line:84% People sort of hope not, and it looks 01:05:02.642 --> 01:05:04.460 align:middle line:90% like the answer might be no. 01:05:04.460 --> 01:05:06.770 align:middle line:84% We won't look at any of their relevant cases today. 01:05:06.770 --> 01:05:08.990 align:middle line:84% So movement A is going to be shorter than movement B 01:05:08.990 --> 01:05:11.360 align:middle line:84% if the path of A contains a smaller set of nodes 01:05:11.360 --> 01:05:13.850 align:middle line:84% than the path of B. I'll show you some trees that 01:05:13.850 --> 01:05:15.820 align:middle line:90% will hopefully make that clear. 01:05:15.820 --> 01:05:18.790 align:middle line:84% So let me show you a case where shortest move is useful, 01:05:18.790 --> 01:05:21.220 align:middle line:84% something called the head movement constraint. 01:05:21.220 --> 01:05:23.020 align:middle line:84% It was invented by Lisa Travis, who's 01:05:23.020 --> 01:05:26.500 align:middle line:84% a syntactician who now teaches at McGill University up 01:05:26.500 --> 01:05:28.300 align:middle line:90% in Montreal. 01:05:28.300 --> 01:05:30.040 align:middle line:84% The head movement constraint says this. 01:05:30.040 --> 01:05:31.373 align:middle line:90% Here's a case of head movement-- 01:05:31.373 --> 01:05:32.950 align:middle line:90% I showed it to you before-- 01:05:32.950 --> 01:05:35.770 align:middle line:84% where we've taken what's in T and we've moved it into C. 01:05:35.770 --> 01:05:38.380 align:middle line:84% So "will Mary type novels," we've taken the auxiliary 01:05:38.380 --> 01:05:40.210 align:middle line:90% and we've moved it into C. 01:05:40.210 --> 01:05:42.880 align:middle line:84% Suppose instead we were to take the verb 01:05:42.880 --> 01:05:46.010 align:middle line:84% and move it into C. Well, we would end up 01:05:46.010 --> 01:05:48.380 align:middle line:84% with "type Mary will novels," which 01:05:48.380 --> 01:05:53.090 align:middle line:84% is not the way you ask yes-no questions in English, no. 01:05:53.090 --> 01:05:54.890 align:middle line:84% And in fact, this generally seems 01:05:54.890 --> 01:05:58.110 align:middle line:84% to be true, that if you are going to do head movement-- 01:05:58.110 --> 01:06:00.350 align:middle line:84% so you're going to move some head into C-- 01:06:00.350 --> 01:06:02.930 align:middle line:84% the head that you move is the higher head. 01:06:02.930 --> 01:06:03.650 align:middle line:90% It's "will." 01:06:03.650 --> 01:06:04.820 align:middle line:90% It's not the lower head. 01:06:04.820 --> 01:06:07.735 align:middle line:90% It's not V. 01:06:07.735 --> 01:06:11.650 align:middle line:84% The path-- so if we're talking about this in terms of paths, 01:06:11.650 --> 01:06:15.910 align:middle line:84% the path from I to C-- sorry, from T to C. 01:06:15.910 --> 01:06:18.070 align:middle line:84% I have to go through and get rid of these I's. 01:06:18.070 --> 01:06:23.650 align:middle line:84% This I, again, is an old name for T. The path from T, from 01:06:23.650 --> 01:06:27.160 align:middle line:84% "will," up into C consists of T bar and TP. 01:06:27.160 --> 01:06:29.290 align:middle line:84% Those are the nodes that dominate T and don't 01:06:29.290 --> 01:06:30.760 align:middle line:90% dominate C. 01:06:30.760 --> 01:06:33.100 align:middle line:84% But the path from V to C consists 01:06:33.100 --> 01:06:35.890 align:middle line:90% of VP and T bar and TP. 01:06:35.890 --> 01:06:36.790 align:middle line:90% Do people see that? 01:06:36.790 --> 01:06:38.207 align:middle line:84% Should I draw that tree again down 01:06:38.207 --> 01:06:41.030 align:middle line:84% here, and I can circle nodes and we can look at them? 01:06:41.030 --> 01:06:43.030 align:middle line:84% So when we're talking about paths-- 01:06:43.030 --> 01:06:47.590 align:middle line:84% again, this is one way of measuring lengths of movement 01:06:47.590 --> 01:06:51.490 align:middle line:84% operations-- the path from V to C 01:06:51.490 --> 01:06:54.130 align:middle line:84% is the set of nodes that dominate V, that don't dominate 01:06:54.130 --> 01:06:58.030 align:middle line:84% C. So those nodes are the VP, which immediately dominates A, 01:06:58.030 --> 01:07:00.430 align:middle line:84% and the T bar, which dominates the VP, 01:07:00.430 --> 01:07:04.510 align:middle line:84% and the T which dominates the T bar, but not the CP, which also 01:07:04.510 --> 01:07:06.520 align:middle line:84% dominates the C. So we're looking just 01:07:06.520 --> 01:07:09.160 align:middle line:84% at the nodes which dominate the place the movement started 01:07:09.160 --> 01:07:11.480 align:middle line:84% and don't dominate the place that it lands. 01:07:11.480 --> 01:07:14.380 align:middle line:90% So that's the path from V to C. 01:07:14.380 --> 01:07:16.960 align:middle line:84% And that path is a superset of the path for movement 01:07:16.960 --> 01:07:19.750 align:middle line:84% from T to C, because VP is in that path, 01:07:19.750 --> 01:07:23.320 align:middle line:84% but it's not in the path from T to C. Does everybody see that? 01:07:23.320 --> 01:07:24.860 align:middle line:90% Is that clear? 01:07:24.860 --> 01:07:25.360 align:middle line:90% OK. 01:07:25.360 --> 01:07:29.110 align:middle line:84% We're mostly going to know short when we see it, right? 01:07:29.110 --> 01:07:31.203 align:middle line:84% So another way to say all of this 01:07:31.203 --> 01:07:32.620 align:middle line:84% would have been to say, hey, look, 01:07:32.620 --> 01:07:35.770 align:middle line:84% the arrow that connects T to C is shorter than the arrow that 01:07:35.770 --> 01:07:37.750 align:middle line:90% connects V to C. There, right? 01:07:37.750 --> 01:07:43.360 align:middle line:84% Or, hey, look, TC commands V. So every node 01:07:43.360 --> 01:07:46.510 align:middle line:84% that dominates T dominates V. That would have done it too. 01:07:46.510 --> 01:07:48.833 align:middle line:84% There are various ways to do it, but this talk 01:07:48.833 --> 01:07:51.250 align:middle line:84% in terms of paths, this is one way people do it sometimes. 01:07:51.250 --> 01:07:54.560 align:middle line:90% 01:07:54.560 --> 01:07:57.655 align:middle line:84% So head movement constraint is one of the sub cases. 01:07:57.655 --> 01:08:00.030 align:middle line:84% So the head movement constraint was invented decades ago. 01:08:00.030 --> 01:08:04.400 align:middle line:84% It's one of the cases that was folded into this general notion 01:08:04.400 --> 01:08:05.750 align:middle line:90% of shortest move. 01:08:05.750 --> 01:08:08.000 align:middle line:84% When you have a choice between things you could move-- 01:08:08.000 --> 01:08:10.160 align:middle line:84% in this case, you're choosing between T and V-- 01:08:10.160 --> 01:08:13.310 align:middle line:84% you have to choose T. You can't choose V. It's what the head 01:08:13.310 --> 01:08:15.400 align:middle line:90% movement constraint says. 01:08:15.400 --> 01:08:18.050 align:middle line:90% Another example. 01:08:18.050 --> 01:08:20.439 align:middle line:84% We talked a little bit about multiple-wh questions, 01:08:20.439 --> 01:08:22.750 align:middle line:84% questions like "Who bought what?" 01:08:22.750 --> 01:08:24.700 align:middle line:90% And speakers vary. 01:08:24.700 --> 01:08:27.700 align:middle line:84% Actually, I'd be interested to hear how you guys feel-- 01:08:27.700 --> 01:08:28.479 align:middle line:90% how you all feel. 01:08:28.479 --> 01:08:34.260 align:middle line:84% I prefer "Who bought what?" to "What did who buy?" 01:08:34.260 --> 01:08:36.990 align:middle line:90% or "What did you give to whom?" 01:08:36.990 --> 01:08:40.720 align:middle line:84% sounds better to me than "Who did you give what to?" 01:08:40.720 --> 01:08:43.100 align:middle line:84% Do people agree with me about that? 01:08:43.100 --> 01:08:45.810 align:middle line:90% Does that sound true? 01:08:45.810 --> 01:08:47.630 align:middle line:84% I saw some of you nodding encouragingly 01:08:47.630 --> 01:08:49.208 align:middle line:90% as I was doing this. 01:08:49.208 --> 01:08:50.750 align:middle line:84% Maybe you just wanted to be friendly. 01:08:50.750 --> 01:08:51.620 align:middle line:90% I appreciate that. 01:08:51.620 --> 01:08:52.520 align:middle line:90% Yeah? 01:08:52.520 --> 01:08:54.545 align:middle line:84% AUDIENCE: I prefer the first options, 01:08:54.545 --> 01:08:56.420 align:middle line:84% but I don't think that the second options are 01:08:56.420 --> 01:08:57.319 align:middle line:90% ungrammatical. 01:08:57.319 --> 01:08:58.310 align:middle line:90% NORVIN RICHARDS: Yeah. 01:08:58.310 --> 01:09:00.560 align:middle line:84% So this is different from the head movement constraint 01:09:00.560 --> 01:09:01.649 align:middle line:90% in at least that regard. 01:09:01.649 --> 01:09:03.899 align:middle line:84% The head movement constraint-- if you violate the head 01:09:03.899 --> 01:09:06.066 align:middle line:84% movement constraint, you've done something very bad. 01:09:06.066 --> 01:09:08.300 align:middle line:84% It's no longer clear what you meant, yeah. 01:09:08.300 --> 01:09:11.149 align:middle line:84% Whereas if I say, "What did who buy?" 01:09:11.149 --> 01:09:12.359 align:middle line:90% people will look at me funny. 01:09:12.359 --> 01:09:13.910 align:middle line:84% But people look at me funny anyway. 01:09:13.910 --> 01:09:16.760 align:middle line:84% That's kind of what life is like, yeah? 01:09:16.760 --> 01:09:19.390 align:middle line:90% 01:09:19.390 --> 01:09:21.130 align:middle line:84% Still, I also agree with Kateryna 01:09:21.130 --> 01:09:23.020 align:middle line:84% that the first examples are better 01:09:23.020 --> 01:09:24.770 align:middle line:90% than the second examples. 01:09:24.770 --> 01:09:28.569 align:middle line:84% This is sometimes called superiority, the phenomenon 01:09:28.569 --> 01:09:30.760 align:middle line:84% that if you have two wh-phrases and you're 01:09:30.760 --> 01:09:33.130 align:middle line:84% trying to pick which one to move, there's a preference. 01:09:33.130 --> 01:09:35.380 align:middle line:84% As Kateryna, I think, accurately says, 01:09:35.380 --> 01:09:37.279 align:middle line:84% it's not a hugely strong preference. 01:09:37.279 --> 01:09:41.770 align:middle line:84% But it's a preference for moving the higher one, the first one. 01:09:41.770 --> 01:09:44.029 align:middle line:84% And this is kind of like the head movement constraint, 01:09:44.029 --> 01:09:46.750 align:middle line:84% and people have subsumed it under shortest move. 01:09:46.750 --> 01:09:48.740 align:middle line:84% You've got a choice between two things to move. 01:09:48.740 --> 01:09:51.310 align:middle line:84% You ought to move the higher one, the one where for which 01:09:51.310 --> 01:09:52.840 align:middle line:90% the movement will be shorter. 01:09:52.840 --> 01:09:53.680 align:middle line:90% Yeah? 01:09:53.680 --> 01:09:55.472 align:middle line:84% AUDIENCE: Could this be related in some way 01:09:55.472 --> 01:09:58.570 align:middle line:84% to what the most important part of the sentence is? 01:09:58.570 --> 01:10:00.580 align:middle line:84% Like, if you hear someone say something 01:10:00.580 --> 01:10:02.330 align:middle line:84% like "We're buying something," and they're 01:10:02.330 --> 01:10:05.260 align:middle line:84% buying this really crazy thing, you 01:10:05.260 --> 01:10:07.110 align:middle line:90% could say, "What did you buy?" 01:10:07.110 --> 01:10:08.200 align:middle line:90% NORVIN RICHARDS: Ah, yes. 01:10:08.200 --> 01:10:11.230 align:middle line:90% 01:10:11.230 --> 01:10:11.890 align:middle line:90% Yeah. 01:10:11.890 --> 01:10:12.520 align:middle line:90% Let's see now. 01:10:12.520 --> 01:10:17.690 align:middle line:90% 01:10:17.690 --> 01:10:19.050 align:middle line:90% Huh. 01:10:19.050 --> 01:10:21.468 align:middle line:84% I'm sorry, I'm going to do an old professor's trick 01:10:21.468 --> 01:10:23.760 align:middle line:84% and talk about something different from what you talked 01:10:23.760 --> 01:10:27.278 align:middle line:84% about, because what you said reminded me of something else. 01:10:27.278 --> 01:10:29.070 align:middle line:84% And then I'm going to talk about that thing 01:10:29.070 --> 01:10:30.300 align:middle line:84% hopefully long enough so that you'll 01:10:30.300 --> 01:10:31.425 align:middle line:90% forget your original point. 01:10:31.425 --> 01:10:32.670 align:middle line:90% That's the goal. 01:10:32.670 --> 01:10:35.910 align:middle line:84% So your job is to stop me from succeeding at this. 01:10:35.910 --> 01:10:39.090 align:middle line:84% But what I thought you were going to say went like this. 01:10:39.090 --> 01:10:41.770 align:middle line:84% There is a kind of wh-question-- we've talked about it. 01:10:41.770 --> 01:10:45.510 align:middle line:84% It's called an echo question, where I just 01:10:45.510 --> 01:10:47.610 align:middle line:84% repeat what you said, except that I 01:10:47.610 --> 01:10:50.130 align:middle line:84% substitute in a wh-phrase which is in-situ. 01:10:50.130 --> 01:10:53.880 align:middle line:84% So if you say "Mary just bought a motorcycle," I can say, 01:10:53.880 --> 01:10:55.830 align:middle line:90% "Mary bought what?" 01:10:55.830 --> 01:10:58.530 align:middle line:84% And similarly, if you say, "Mary just bought 01:10:58.530 --> 01:11:01.290 align:middle line:84% a motorcycle," I can say, "Who bought a motorcycle?" 01:11:01.290 --> 01:11:03.960 align:middle line:84% And because that's at the beginning of the sentence, 01:11:03.960 --> 01:11:07.000 align:middle line:84% it's hard to know whether that's an echo question or not. 01:11:07.000 --> 01:11:10.440 align:middle line:84% But at least it's possible that it is. 01:11:10.440 --> 01:11:12.870 align:middle line:84% And I can do that either because it's hard to believe-- 01:11:12.870 --> 01:11:15.412 align:middle line:84% there are various people I could believe bought a motorcycle, 01:11:15.412 --> 01:11:18.600 align:middle line:84% but not Mary, we all know Mary would never buy a motorcycle-- 01:11:18.600 --> 01:11:20.268 align:middle line:84% or I could do it because I didn't 01:11:20.268 --> 01:11:21.810 align:middle line:84% hear the first word of your sentence, 01:11:21.810 --> 01:11:24.270 align:middle line:84% like the phone connection is bad or we're 01:11:24.270 --> 01:11:26.772 align:middle line:84% talking on the T or whatever, and there was a loud noise 01:11:26.772 --> 01:11:29.230 align:middle line:84% right when you said "Mary," and so I didn't hear that part. 01:11:29.230 --> 01:11:31.665 align:middle line:84% And I can say, "Who bought a motorcycle?" 01:11:31.665 --> 01:11:34.290 align:middle line:84% Not because I'm astonished, but because I missed the first part 01:11:34.290 --> 01:11:36.660 align:middle line:90% of your sentence. 01:11:36.660 --> 01:11:39.990 align:middle line:84% And "What did who buy?" that string, 01:11:39.990 --> 01:11:44.460 align:middle line:84% I think it's possible to have an echo question with wh-movement 01:11:44.460 --> 01:11:45.450 align:middle line:90% in it. 01:11:45.450 --> 01:11:50.670 align:middle line:84% So if you've just asked me, "What did Mary buy?" 01:11:50.670 --> 01:11:55.390 align:middle line:84% and I didn't hear you say "Mary" because the line was bad. 01:11:55.390 --> 01:11:57.660 align:middle line:84% So you said, "What did [IMITATING STATIC] buy?" 01:11:57.660 --> 01:12:01.370 align:middle line:90% I can say, "What did who buy?" 01:12:01.370 --> 01:12:05.060 align:middle line:84% Which resembles what you were talking about, 01:12:05.060 --> 01:12:06.510 align:middle line:90% but it's not quite the same. 01:12:06.510 --> 01:12:11.540 align:middle line:84% You wanted-- so that's a place where my two wh-phrases have 01:12:11.540 --> 01:12:14.150 align:middle line:90% different goals. 01:12:14.150 --> 01:12:16.820 align:middle line:90% When I say, "What did who buy?" 01:12:16.820 --> 01:12:18.830 align:middle line:84% I'm really asking you the question "who," 01:12:18.830 --> 01:12:20.893 align:middle line:90% and my "what" is quoting you. 01:12:20.893 --> 01:12:23.060 align:middle line:84% What you said was "What did [IMITATING STATIC] buy?" 01:12:23.060 --> 01:12:24.920 align:middle line:90% right. 01:12:24.920 --> 01:12:27.290 align:middle line:84% That's not what you said, but that's what I heard. 01:12:27.290 --> 01:12:30.080 align:middle line:84% And so I'm doing an echo question 01:12:30.080 --> 01:12:31.380 align:middle line:90% about your wh-question. 01:12:31.380 --> 01:12:31.880 align:middle line:90% Yeah. 01:12:31.880 --> 01:12:34.430 align:middle line:90% My real question is about "who." 01:12:34.430 --> 01:12:39.770 align:middle line:84% Your idea was that if I ask, "What did who buy?" 01:12:39.770 --> 01:12:40.960 align:middle line:90% "What did who buy?" 01:12:40.960 --> 01:12:42.710 align:middle line:84% No, your question was, "What did who buy?" 01:12:42.710 --> 01:12:43.490 align:middle line:90% What was your question? 01:12:43.490 --> 01:12:44.900 align:middle line:84% AUDIENCE: Like, if you're more interested in the what 01:12:44.900 --> 01:12:45.710 align:middle line:90% than who. 01:12:45.710 --> 01:12:46.460 align:middle line:84% NORVIN RICHARDS: If I'm more interested in the what 01:12:46.460 --> 01:12:47.480 align:middle line:90% than the who. 01:12:47.480 --> 01:12:51.080 align:middle line:90% 01:12:51.080 --> 01:12:54.210 align:middle line:90% So this is related-- 01:12:54.210 --> 01:12:54.710 align:middle line:90% yeah. 01:12:54.710 --> 01:12:57.530 align:middle line:90% 01:12:57.530 --> 01:12:59.050 align:middle line:90% This is related to other things. 01:12:59.050 --> 01:13:01.490 align:middle line:84% Let me see, what's a good example of this? 01:13:01.490 --> 01:13:07.985 align:middle line:84% It's a thing about wh-answers to wh-questions. 01:13:07.985 --> 01:13:09.985 align:middle line:84% if I ask you a question like, "Who bought what?" 01:13:09.985 --> 01:13:12.730 align:middle line:90% 01:13:12.730 --> 01:13:16.840 align:middle line:84% I want a complete list of people. 01:13:16.840 --> 01:13:20.980 align:middle line:84% I want my answer organized as a complete list of people, 01:13:20.980 --> 01:13:25.960 align:middle line:84% together with the things that they bought. 01:13:25.960 --> 01:13:31.690 align:middle line:84% Whereas if I make myself ask you, "What did who buy?" 01:13:31.690 --> 01:13:35.590 align:middle line:84% what I want is a complete list of things 01:13:35.590 --> 01:13:38.290 align:middle line:84% organized by things together with the people who 01:13:38.290 --> 01:13:39.400 align:middle line:90% bought them. 01:13:39.400 --> 01:13:42.280 align:middle line:84% And there's some debate about whether-- so 01:13:42.280 --> 01:13:44.860 align:middle line:84% under normal circumstances, wh-questions 01:13:44.860 --> 01:13:47.050 align:middle line:90% are supposed to be exhaustive. 01:13:47.050 --> 01:13:49.120 align:middle line:84% Answers are supposed to be exhaustive, that is. 01:13:49.120 --> 01:13:55.450 align:middle line:84% If I ask you, "What did you eat yesterday?" 01:13:55.450 --> 01:13:59.110 align:middle line:84% you're supposed to give me a more or less complete list. 01:13:59.110 --> 01:14:02.290 align:middle line:84% If you leave out all the desserts and I'm your doctor, 01:14:02.290 --> 01:14:04.180 align:middle line:84% then you are you're doing something wrong. 01:14:04.180 --> 01:14:07.120 align:middle line:84% You're supposed to tell me everything. 01:14:07.120 --> 01:14:09.550 align:middle line:84% There are circumstances under which it clearly isn't 01:14:09.550 --> 01:14:10.790 align:middle line:90% supposed to be exhaustive. 01:14:10.790 --> 01:14:13.180 align:middle line:84% So if I ask you a question like, "Where can I 01:14:13.180 --> 01:14:15.192 align:middle line:84% buy a foreign newspaper around here?" 01:14:15.192 --> 01:14:16.900 align:middle line:84% you don't have to give me a complete list 01:14:16.900 --> 01:14:19.233 align:middle line:84% of the places I can buy a foreign newspaper around here. 01:14:19.233 --> 01:14:22.780 align:middle line:84% You just have to tell me a place where I can go. 01:14:22.780 --> 01:14:24.700 align:middle line:90% "Harvard Square," you can say. 01:14:24.700 --> 01:14:26.680 align:middle line:90% Maybe that's where I can go. 01:14:26.680 --> 01:14:29.470 align:middle line:84% And with multiple wh-questions, there's 01:14:29.470 --> 01:14:32.410 align:middle line:84% some debate about whether the exhaustivity requirement, 01:14:32.410 --> 01:14:34.960 align:middle line:84% to the extent that it ever holds with wh-questions, 01:14:34.960 --> 01:14:38.140 align:middle line:84% maybe only holds of the one that you move and not of the one 01:14:38.140 --> 01:14:39.640 align:middle line:90% that you've left in situ. 01:14:39.640 --> 01:14:42.220 align:middle line:84% That's a position people have held anyway, that there's 01:14:42.220 --> 01:14:44.480 align:middle line:90% a difference between the two. 01:14:44.480 --> 01:14:46.210 align:middle line:84% And so your question about importance 01:14:46.210 --> 01:14:47.740 align:middle line:90% might be getting at that, right? 01:14:47.740 --> 01:14:50.080 align:middle line:84% It's like, I want this one completely exhausted, 01:14:50.080 --> 01:14:51.010 align:middle line:90% and then-- 01:14:51.010 --> 01:14:53.410 align:middle line:84% yes, I want every member of it connected to something 01:14:53.410 --> 01:14:54.820 align:middle line:84% in the other group, but you don't 01:14:54.820 --> 01:14:58.240 align:middle line:84% have to be exhaustive with the pairings. 01:14:58.240 --> 01:15:00.738 align:middle line:90% That might be why-- 01:15:00.738 --> 01:15:02.530 align:middle line:84% I think this was Kateryna's comment earlier 01:15:02.530 --> 01:15:04.420 align:middle line:84% that the violations of this are not 01:15:04.420 --> 01:15:07.003 align:middle line:84% as strong as the violations of the head movement constraint. 01:15:07.003 --> 01:15:08.920 align:middle line:84% And it might have to do with things like this, 01:15:08.920 --> 01:15:11.230 align:middle line:84% that there are extraordinary circumstances under which 01:15:11.230 --> 01:15:14.170 align:middle line:84% these effects might be overwhelmed by your desire 01:15:14.170 --> 01:15:15.520 align:middle line:90% to have-- 01:15:15.520 --> 01:15:16.812 align:middle line:90% to prioritize one or the other. 01:15:16.812 --> 01:15:18.270 align:middle line:84% Whereas with head movement, there's 01:15:18.270 --> 01:15:19.540 align:middle line:90% nothing like that going on. 01:15:19.540 --> 01:15:22.803 align:middle line:90% It's just got to move ahead. 01:15:22.803 --> 01:15:24.970 align:middle line:84% One of the cases that people talk about when they're 01:15:24.970 --> 01:15:26.980 align:middle line:84% talking about exhaustivity is cases 01:15:26.980 --> 01:15:30.740 align:middle line:84% where you know that there's only a single pair. 01:15:30.740 --> 01:15:37.840 align:middle line:84% So if I'm eating lunch with a colleague, 01:15:37.840 --> 01:15:39.940 align:middle line:84% if I've invited a colleague to have lunch with me 01:15:39.940 --> 01:15:41.380 align:middle line:84% and we go and have lunch, and then 01:15:41.380 --> 01:15:44.497 align:middle line:84% I try to pay the bill and my colleague 01:15:44.497 --> 01:15:46.330 align:middle line:84% is arguing with me about whether-- about who 01:15:46.330 --> 01:15:50.800 align:middle line:84% gets to pay, I might say, "Well, who invited who?" 01:15:50.800 --> 01:15:52.090 align:middle line:90% Right? 01:15:52.090 --> 01:15:55.120 align:middle line:84% I'm not going to say, "Who did who invite?" 01:15:55.120 --> 01:15:56.620 align:middle line:90% And maybe it's-- I think. 01:15:56.620 --> 01:15:59.110 align:middle line:84% That's a case where the judgment is pretty clear. 01:15:59.110 --> 01:16:01.960 align:middle line:84% And I think that might be a case where there's 01:16:01.960 --> 01:16:03.550 align:middle line:90% no question of exhaustivity. 01:16:03.550 --> 01:16:04.930 align:middle line:84% There are only two people, right? 01:16:04.930 --> 01:16:07.240 align:middle line:84% It's either I invited you or you invited me. 01:16:07.240 --> 01:16:09.940 align:middle line:84% Those are the only possible answers. 01:16:09.940 --> 01:16:13.090 align:middle line:84% And so maybe that's why the judgment, I think, 01:16:13.090 --> 01:16:14.990 align:middle line:84% in that kind of case gets stronger. 01:16:14.990 --> 01:16:15.490 align:middle line:90% Yeah. 01:16:15.490 --> 01:16:17.350 align:middle line:84% Sorry, did I ever manage to answer your question 01:16:17.350 --> 01:16:17.980 align:middle line:90% at any point of that? 01:16:17.980 --> 01:16:18.530 align:middle line:90% Yes, good. 01:16:18.530 --> 01:16:19.030 align:middle line:90% Yeah. 01:16:19.030 --> 01:16:21.590 align:middle line:84% This is all-- you're opening a can of worms, 01:16:21.590 --> 01:16:23.020 align:middle line:90% which I will now close. 01:16:23.020 --> 01:16:26.200 align:middle line:84% There's lots of interesting work on exactly this problem, yeah. 01:16:26.200 --> 01:16:28.630 align:middle line:84% And we are just about out of time. 01:16:28.630 --> 01:16:31.130 align:middle line:84% Yes, we're out of time, because this one will take too long. 01:16:31.130 --> 01:16:33.010 align:middle line:84% So we'll do more shortest move next time. 01:16:33.010 --> 01:16:34.330 align:middle line:90% We'll talk more about this. 01:16:34.330 --> 01:16:34.870 align:middle line:90% Cool. 01:16:34.870 --> 01:16:36.420 align:middle line:90% Thanks. 01:16:36.420 --> 01:16:41.000 align:middle line:90%