WEBVTT 00:00:00.000 --> 00:00:01.996 align:middle line:90% [SQUEAKING] 00:00:01.996 --> 00:00:03.992 align:middle line:90% [RUSTLING] 00:00:03.992 --> 00:00:05.489 align:middle line:90% [CLICKING] 00:00:05.489 --> 00:00:10.978 align:middle line:90% 00:00:10.978 --> 00:00:12.490 align:middle line:90% NORVIN RICHARDS: Sorry. 00:00:12.490 --> 00:00:16.127 align:middle line:90% So here's syntax five. 00:00:16.127 --> 00:00:17.710 align:middle line:84% If you're wondering whether we're ever 00:00:17.710 --> 00:00:21.550 align:middle line:84% going to stop doing syntax, we are, but I'm not sure when. 00:00:21.550 --> 00:00:25.900 align:middle line:84% I think we have at least one or two more days of syntax. 00:00:25.900 --> 00:00:27.910 align:middle line:84% So if you're looking at the syllabus 00:00:27.910 --> 00:00:31.480 align:middle line:84% to try to figure out where we are, just don't do that. 00:00:31.480 --> 00:00:33.820 align:middle line:84% I'll try to update the syllabus soon 00:00:33.820 --> 00:00:39.530 align:middle line:84% to give you some idea of what's going to happen in the future. 00:00:39.530 --> 00:00:42.850 align:middle line:84% Last time, we got started on something 00:00:42.850 --> 00:00:46.010 align:middle line:84% that I promised you at the beginning of the semester. 00:00:46.010 --> 00:00:46.510 align:middle line:90% Sorry. 00:00:46.510 --> 00:00:48.177 align:middle line:84% I'm just going to close this real quick. 00:00:48.177 --> 00:00:50.225 align:middle line:90% 00:00:50.225 --> 00:00:51.850 align:middle line:84% I said at the beginning of the semester 00:00:51.850 --> 00:00:54.225 align:middle line:84% that one of the kinds of things we were going to find out 00:00:54.225 --> 00:00:58.023 align:middle line:84% was that, although there were lots kinds of languages 00:00:58.023 --> 00:00:59.440 align:middle line:84% out there in the world and we were 00:00:59.440 --> 00:01:02.080 align:middle line:84% going to have a chance to look at some of them, 00:01:02.080 --> 00:01:05.700 align:middle line:84% that one of the things that we find as we study these-- 00:01:05.700 --> 00:01:07.700 align:middle line:84% I turned out not to be able to close that door-- 00:01:07.700 --> 00:01:09.940 align:middle line:90% Let me try that again. 00:01:09.940 --> 00:01:13.840 align:middle line:84% One of the things that we find as we 00:01:13.840 --> 00:01:17.080 align:middle line:84% study languages of the world is that, although there 00:01:17.080 --> 00:01:18.830 align:middle line:90% are many languages out there-- 00:01:18.830 --> 00:01:22.540 align:middle line:84% oh, well-- I've been defeated by a door. 00:01:22.540 --> 00:01:23.798 align:middle line:90% This is very sad. 00:01:23.798 --> 00:01:25.840 align:middle line:84% That although there are many languages out there, 00:01:25.840 --> 00:01:28.150 align:middle line:84% there are kinds of languages that are easy to imagine 00:01:28.150 --> 00:01:30.590 align:middle line:90% but that don't, in fact, exist. 00:01:30.590 --> 00:01:35.690 align:middle line:84% And there's a hypothesis about what's going on, 00:01:35.690 --> 00:01:39.010 align:middle line:84% which is that part of being a human being 00:01:39.010 --> 00:01:42.370 align:middle line:84% is having the kind of mind that can construct language 00:01:42.370 --> 00:01:43.870 align:middle line:90% in some ways, but not others. 00:01:43.870 --> 00:01:47.920 align:middle line:84% That is, given a certain set of data, 00:01:47.920 --> 00:01:51.550 align:middle line:84% the human mind jumps to certain conclusions but not others 00:01:51.550 --> 00:01:54.525 align:middle line:84% about this system that's underlying those data. 00:01:54.525 --> 00:01:56.650 align:middle line:84% And one of the things that we're trying to find out 00:01:56.650 --> 00:01:58.990 align:middle line:84% as we study linguistics is about what 00:01:58.990 --> 00:02:01.340 align:middle line:84% it is about human minds that makes them work this way. 00:02:01.340 --> 00:02:03.580 align:middle line:84% What are the rules that our minds 00:02:03.580 --> 00:02:07.680 align:middle line:84% are using to make sense of the limited data that we have? 00:02:07.680 --> 00:02:10.630 align:middle line:84% We ran very quickly through one argument of that kind 00:02:10.630 --> 00:02:12.873 align:middle line:84% last time because we were running low on time, 00:02:12.873 --> 00:02:15.040 align:middle line:84% and I thought, maybe we have time for this argument. 00:02:15.040 --> 00:02:18.220 align:middle line:84% And probably, looking back, we probably didn't. 00:02:18.220 --> 00:02:20.770 align:middle line:84% If you are looking at the-- if you 00:02:20.770 --> 00:02:22.540 align:middle line:84% get a chance to look at the video of that, 00:02:22.540 --> 00:02:24.290 align:middle line:84% if anybody wants to, you should look at it 00:02:24.290 --> 00:02:28.400 align:middle line:84% on slowed down speed to see if you can get it to make sense. 00:02:28.400 --> 00:02:30.430 align:middle line:84% So I was going to go through that argument again 00:02:30.430 --> 00:02:32.530 align:middle line:84% and some other arguments a little more. 00:02:32.530 --> 00:02:34.150 align:middle line:84% That's going to be a theme for today. 00:02:34.150 --> 00:02:38.390 align:middle line:84% I will show you some areas in which languages vary, 00:02:38.390 --> 00:02:41.480 align:middle line:84% and then we will imagine kinds of languages, 00:02:41.480 --> 00:02:44.080 align:middle line:84% and we'll discover that those kinds that are easy to imagine 00:02:44.080 --> 00:02:46.160 align:middle line:90% don't happen. 00:02:46.160 --> 00:02:49.330 align:middle line:90% So we're going to look at that. 00:02:49.330 --> 00:02:50.770 align:middle line:90% This is 24.900. 00:02:50.770 --> 00:02:52.870 align:middle line:84% If you go on and take more linguistics classes, 00:02:52.870 --> 00:02:55.810 align:middle line:84% you will see more phenomena of this kind and more theories 00:02:55.810 --> 00:02:57.950 align:middle line:84% about why languages behave this way. 00:02:57.950 --> 00:03:00.760 align:middle line:84% But trying to understand why languages behave this way 00:03:00.760 --> 00:03:03.970 align:middle line:84% is one of the central occupations of linguistics, 00:03:03.970 --> 00:03:07.730 align:middle line:84% trying to understand basically what human minds are like. 00:03:07.730 --> 00:03:10.250 align:middle line:84% So last time I showed you the slide-- 00:03:10.250 --> 00:03:12.470 align:middle line:84% this was a slide for the question, "What will Mary 00:03:12.470 --> 00:03:13.670 align:middle line:90% write?" 00:03:13.670 --> 00:03:16.340 align:middle line:84% I was trying to convince you that, 00:03:16.340 --> 00:03:20.510 align:middle line:84% in a sentence like this where we think that "write" selects 00:03:20.510 --> 00:03:21.230 align:middle line:90% an object-- 00:03:21.230 --> 00:03:24.440 align:middle line:84% it has the option of being transitive-- 00:03:24.440 --> 00:03:26.900 align:middle line:84% you can say Mary is writing a novel. 00:03:26.900 --> 00:03:30.140 align:middle line:84% And we thought that when you had a selection relation 00:03:30.140 --> 00:03:32.960 align:middle line:84% between the head and some phrase, 00:03:32.960 --> 00:03:36.220 align:middle line:84% that the phrase needed to be the sister to the head. 00:03:36.220 --> 00:03:38.720 align:middle line:84% At least, if there was only one thing that it was selecting. 00:03:38.720 --> 00:03:41.305 align:middle line:84% We spent a lot of happy time thinking 00:03:41.305 --> 00:03:42.680 align:middle line:84% about what happens when a head is 00:03:42.680 --> 00:03:43.940 align:middle line:90% selecting more than one thing. 00:03:43.940 --> 00:03:48.410 align:middle line:84% And I promised you that we would talk more about that later. 00:03:48.410 --> 00:03:50.600 align:middle line:84% But that promise did not apply to today. 00:03:50.600 --> 00:03:51.920 align:middle line:90% So today, yeah. 00:03:51.920 --> 00:03:53.630 align:middle line:84% If a head is selecting only one thing, 00:03:53.630 --> 00:03:55.250 align:middle line:84% that thing needs to be the sister. 00:03:55.250 --> 00:03:58.010 align:middle line:84% And then we looked at questions like this, 00:03:58.010 --> 00:04:01.710 align:middle line:84% and I tried to convince you to be upset. 00:04:01.710 --> 00:04:04.172 align:middle line:84% So the word "what," which seems to be the thing that's 00:04:04.172 --> 00:04:05.630 align:middle line:84% being selected by "write," it seems 00:04:05.630 --> 00:04:07.010 align:middle line:90% to be the thing that's getting-- 00:04:07.010 --> 00:04:09.170 align:middle line:90% that is the object of "write"-- 00:04:09.170 --> 00:04:10.850 align:middle line:90% isn't anywhere near "write." 00:04:10.850 --> 00:04:12.530 align:middle line:84% What I said was there is a standard way 00:04:12.530 --> 00:04:14.030 align:middle line:84% to deal with this problem, and we're 00:04:14.030 --> 00:04:17.180 align:middle line:84% going to gradually accumulate evidence for this idea. 00:04:17.180 --> 00:04:20.250 align:middle line:84% But for today, I'll just show you the idea. 00:04:20.250 --> 00:04:24.290 align:middle line:84% The idea is that, indeed, what does start off 00:04:24.290 --> 00:04:29.390 align:middle line:84% as the sister of "write" goes right where the projection 00:04:29.390 --> 00:04:30.800 align:middle line:90% principle says it should. 00:04:30.800 --> 00:04:32.270 align:middle line:90% It's merged right away. 00:04:32.270 --> 00:04:33.920 align:middle line:84% As soon as "write" is introduced, 00:04:33.920 --> 00:04:35.420 align:middle line:84% first thing you merge "write" with 00:04:35.420 --> 00:04:37.700 align:middle line:84% is something that "write" is selecting-- its object. 00:04:37.700 --> 00:04:41.660 align:middle line:84% But then there is an operation that obscures that, destroys 00:04:41.660 --> 00:04:44.323 align:middle line:84% that relation, takes that thing which starts off 00:04:44.323 --> 00:04:46.490 align:middle line:84% as the sister of "write" and moves it to this higher 00:04:46.490 --> 00:04:48.620 align:middle line:90% position in the clause. 00:04:48.620 --> 00:04:52.010 align:middle line:84% I said that it moved to the specifying of CP. 00:04:52.010 --> 00:04:54.780 align:middle line:84% That is that position that's there on the tree. 00:04:54.780 --> 00:04:58.250 align:middle line:84% So it becomes the daughter of the CP node. 00:04:58.250 --> 00:05:01.130 align:middle line:84% CP, maybe you remember, is the projection 00:05:01.130 --> 00:05:03.740 align:middle line:84% that's headed by words like "that" and "whether" 00:05:03.740 --> 00:05:07.640 align:middle line:84% in sentences like "I think that it is raining," talking 00:05:07.640 --> 00:05:09.890 align:middle line:84% about that as being the place where we would put words 00:05:09.890 --> 00:05:11.400 align:middle line:90% like that. 00:05:11.400 --> 00:05:14.180 align:middle line:84% Does this all sound familiar and not terribly upsetting? 00:05:14.180 --> 00:05:16.850 align:middle line:84% Is there anything in any of this that makes any of you 00:05:16.850 --> 00:05:19.370 align:middle line:84% want to demand further explanation 00:05:19.370 --> 00:05:22.070 align:middle line:84% or even ask nicely for further explanation? 00:05:22.070 --> 00:05:23.420 align:middle line:90% Anything else like that? 00:05:23.420 --> 00:05:23.960 align:middle line:90% No? 00:05:23.960 --> 00:05:25.980 align:middle line:90% OK. 00:05:25.980 --> 00:05:30.305 align:middle line:90% So one of the-- so wh-movement. 00:05:30.305 --> 00:05:31.680 align:middle line:84% Cross-linguistically very common. 00:05:31.680 --> 00:05:34.110 align:middle line:84% Lots of languages out there that have this operation 00:05:34.110 --> 00:05:37.080 align:middle line:84% that takes these wh-words, these question words 00:05:37.080 --> 00:05:39.150 align:middle line:84% with meanings like "what" and "who" and "where" 00:05:39.150 --> 00:05:43.540 align:middle line:84% and "why" and puts them at the beginning of the sentence. 00:05:43.540 --> 00:05:46.560 align:middle line:84% So hearing examples from English and Tagalog and Finnish. 00:05:46.560 --> 00:05:49.080 align:middle line:90% 00:05:49.080 --> 00:05:52.598 align:middle line:84% The language that you're working on could be a kind of thing 00:05:52.598 --> 00:05:54.390 align:middle line:84% that you could try to find out "How does it 00:05:54.390 --> 00:05:55.590 align:middle line:90% do its wh-questions?" 00:05:55.590 --> 00:05:58.560 align:middle line:84% Because there is cross-linguistic variation-- 00:05:58.560 --> 00:06:00.660 align:middle line:90% we said this last time too-- 00:06:00.660 --> 00:06:02.680 align:middle line:84% with respect to how you form the wh-question. 00:06:02.680 --> 00:06:04.222 align:middle line:84% So there are languages, like English, 00:06:04.222 --> 00:06:06.540 align:middle line:90% which do wh-movement usually. 00:06:06.540 --> 00:06:09.840 align:middle line:84% We talked about special conversational circumstances 00:06:09.840 --> 00:06:11.940 align:middle line:84% under which you might leave a "wh" in-situ. 00:06:11.940 --> 00:06:14.790 align:middle line:84% So there are times when it's possible to say 00:06:14.790 --> 00:06:17.070 align:middle line:90% "Mary wrote what." 00:06:17.070 --> 00:06:19.950 align:middle line:84% But it specifically has to involve either 00:06:19.950 --> 00:06:22.890 align:middle line:84% not having heard you say exactly what Mary wrote 00:06:22.890 --> 00:06:25.200 align:middle line:90% or being amazed. 00:06:25.200 --> 00:06:30.120 align:middle line:84% So you say, "Mary wrote an Estonian novel." 00:06:30.120 --> 00:06:31.620 align:middle line:90% And I say, "Mary wrote what? 00:06:31.620 --> 00:06:35.310 align:middle line:84% I didn't know she spoke Estonian." 00:06:35.310 --> 00:06:37.462 align:middle line:84% But except under those special circumstances, 00:06:37.462 --> 00:06:39.420 align:middle line:84% the normal thing to do with wh-words in English 00:06:39.420 --> 00:06:41.420 align:middle line:84% is to put them at the beginning of the sentence. 00:06:41.420 --> 00:06:44.160 align:middle line:84% There are other languages, like Mandarin or Bafut-- 00:06:44.160 --> 00:06:46.860 align:middle line:84% which is a language of Cameroon-- or Hopi-- 00:06:46.860 --> 00:06:49.290 align:middle line:84% which is a language of the American Southwest-- 00:06:49.290 --> 00:06:51.430 align:middle line:84% many, many languages which don't do that. 00:06:51.430 --> 00:06:53.010 align:middle line:84% They have what's called "wh in-situ." 00:06:53.010 --> 00:06:56.640 align:middle line:84% That is, the wh-phrases just stay where they would normally 00:06:56.640 --> 00:06:59.370 align:middle line:90% be in the language in question. 00:06:59.370 --> 00:07:03.098 align:middle line:84% So Chinese word order is not unlike English word order. 00:07:03.098 --> 00:07:05.640 align:middle line:84% You've got the subject and then the verb and then the object. 00:07:05.640 --> 00:07:08.735 align:middle line:84% And the word for "what" there is sitting right 00:07:08.735 --> 00:07:10.110 align:middle line:84% where an object would normally be 00:07:10.110 --> 00:07:15.120 align:middle line:84% if you wanted to say "Zhangsan bought an Estonian novel" 00:07:15.120 --> 00:07:15.870 align:middle line:90% in Mandarin-- 00:07:15.870 --> 00:07:17.162 align:middle line:90% I have no idea how to say that. 00:07:17.162 --> 00:07:19.410 align:middle line:84% I don't speak Mandarin well at all. 00:07:19.410 --> 00:07:22.080 align:middle line:84% The words for "Estonian novel" would go right after "bought." 00:07:22.080 --> 00:07:24.390 align:middle line:84% It would go right where "what" is. 00:07:24.390 --> 00:07:25.860 align:middle line:90% Yep. 00:07:25.860 --> 00:07:28.590 align:middle line:90% Same deal for Bafut and Hopi. 00:07:28.590 --> 00:07:29.970 align:middle line:90% Hopi word order is different. 00:07:29.970 --> 00:07:31.595 align:middle line:84% The verb is at the end of the sentence, 00:07:31.595 --> 00:07:33.600 align:middle line:84% but the generalization about all these languages 00:07:33.600 --> 00:07:36.000 align:middle line:84% is that the wh-phrase is just staying where it is. 00:07:36.000 --> 00:07:38.973 align:middle line:84% There is no WH-movement going on. 00:07:38.973 --> 00:07:40.890 align:middle line:84% And what I said last time-- and this is true-- 00:07:40.890 --> 00:07:43.560 align:middle line:84% is those seem to be the options, basically. 00:07:43.560 --> 00:07:46.215 align:middle line:84% You either put your wh-phrase at the beginning of the sentence 00:07:46.215 --> 00:07:51.120 align:middle line:84% when you do wh-movement or you leave it where it is. 00:07:51.120 --> 00:07:53.600 align:middle line:84% You could imagine other things that you would do. 00:07:53.600 --> 00:07:55.850 align:middle line:84% You could imagine a language where the wh-phrase would 00:07:55.850 --> 00:07:58.130 align:middle line:84% move to the end of the sentence or a language 00:07:58.130 --> 00:07:59.678 align:middle line:84% where the wh-phrase would have to be 00:07:59.678 --> 00:08:01.220 align:middle line:84% in the middle of the sentence-- you'd 00:08:01.220 --> 00:08:05.240 align:middle line:84% count the words in the sentence and make sure it was evenly 00:08:05.240 --> 00:08:07.100 align:middle line:90% in the middle of the sentence. 00:08:07.100 --> 00:08:10.880 align:middle line:84% There aren't any languages like that. 00:08:10.880 --> 00:08:12.743 align:middle line:90% Now, so here's an example. 00:08:12.743 --> 00:08:14.660 align:middle line:84% There are no languages in which the normal way 00:08:14.660 --> 00:08:16.370 align:middle line:84% to say "who has eaten the cookies" is 00:08:16.370 --> 00:08:17.300 align:middle line:84% "has eaten the cookies who?" where 00:08:17.300 --> 00:08:18.883 align:middle line:84% I've put in all these silly diacritics 00:08:18.883 --> 00:08:20.905 align:middle line:84% to emphasize the fact that I'm making this up. 00:08:20.905 --> 00:08:22.280 align:middle line:84% There are no languages like this. 00:08:22.280 --> 00:08:23.655 align:middle line:90% This is not a real language. 00:08:23.655 --> 00:08:25.220 align:middle line:90% So I'm making up conlang. 00:08:25.220 --> 00:08:25.735 align:middle line:90% Yeah? 00:08:25.735 --> 00:08:27.110 align:middle line:84% AUDIENCE: Are there any languages 00:08:27.110 --> 00:08:30.230 align:middle line:84% in which [? move ?] to the beginning of the sentence 00:08:30.230 --> 00:08:35.534 align:middle line:84% or just the in situ [INAUDIBLE] are they both [INAUDIBLE].. 00:08:35.534 --> 00:08:36.409 align:middle line:90% NORVIN RICHARDS: Yes. 00:08:36.409 --> 00:08:37.130 align:middle line:90% Yes. 00:08:37.130 --> 00:08:39.130 align:middle line:84% Excellent question, and I should have said that. 00:08:39.130 --> 00:08:41.360 align:middle line:84% There are languages that allow you to do either one. 00:08:41.360 --> 00:08:44.030 align:middle line:84% French is a language like that, for example. 00:08:44.030 --> 00:08:49.012 align:middle line:84% And, in fact, it's not uncommon for languages like that to-- 00:08:49.012 --> 00:08:50.720 align:middle line:84% I mean, the first thing to say about them 00:08:50.720 --> 00:08:52.790 align:middle line:90% is that they have both options. 00:08:52.790 --> 00:08:55.580 align:middle line:84% It's not uncommon for one of the options to be more common 00:08:55.580 --> 00:08:57.650 align:middle line:90% or to be less restricted. 00:08:57.650 --> 00:09:01.190 align:middle line:84% So French allows you to do wh-in-situ, but not always. 00:09:01.190 --> 00:09:03.093 align:middle line:84% There are restrictions on it, and people 00:09:03.093 --> 00:09:04.760 align:middle line:84% get very interested in the restrictions, 00:09:04.760 --> 00:09:08.360 align:middle line:84% trying to figure out when do you have to wh-move in French. 00:09:08.360 --> 00:09:09.110 align:middle line:90% But yes. 00:09:09.110 --> 00:09:11.960 align:middle line:84% The short version of the answer to your question is yes. 00:09:11.960 --> 00:09:15.440 align:middle line:90% Other questions about this? 00:09:15.440 --> 00:09:16.390 align:middle line:90% OK. 00:09:16.390 --> 00:09:18.660 align:middle line:84% So I started this off by saying we're 00:09:18.660 --> 00:09:20.660 align:middle line:84% going to spend some time talking about languages 00:09:20.660 --> 00:09:21.540 align:middle line:90% that don't exist. 00:09:21.540 --> 00:09:22.850 align:middle line:84% Here's a language that doesn't exist. 00:09:22.850 --> 00:09:25.225 align:middle line:84% There aren't languages that take wh-phrases and move them 00:09:25.225 --> 00:09:27.150 align:middle line:90% to the end of the sentence. 00:09:27.150 --> 00:09:30.592 align:middle line:84% And if you're syntactician, you want to know, well, why? 00:09:30.592 --> 00:09:32.300 align:middle line:84% Why aren't there any languages like that? 00:09:32.300 --> 00:09:35.540 align:middle line:90% It's not hard to imagine them. 00:09:35.540 --> 00:09:39.990 align:middle line:84% Now, when I am trapped on an airplane 00:09:39.990 --> 00:09:42.338 align:middle line:84% and the person next to me wants to talk to me, 00:09:42.338 --> 00:09:44.130 align:middle line:84% and, for some reason, they want to continue 00:09:44.130 --> 00:09:45.510 align:middle line:84% to talk to me after I tell them that I'm 00:09:45.510 --> 00:09:46.770 align:middle line:90% a theoretical syntactician. 00:09:46.770 --> 00:09:48.510 align:middle line:84% If I tell them this, if I tell them 00:09:48.510 --> 00:09:51.360 align:middle line:84% about this kind of contrast, they sometimes 00:09:51.360 --> 00:09:53.550 align:middle line:84% respond by saying-- sometimes they say, "Oh, yes. 00:09:53.550 --> 00:09:54.717 align:middle line:90% That's very interesting...." 00:09:54.717 --> 00:09:56.580 align:middle line:84% And then they start reading their novel. 00:09:56.580 --> 00:09:59.850 align:middle line:84% But sometimes they respond by saying, yeah, but look. 00:09:59.850 --> 00:10:04.560 align:middle line:84% In a wh-question, the wh-word is the important thing. 00:10:04.560 --> 00:10:06.570 align:middle line:90% So maybe it's just-- 00:10:06.570 --> 00:10:08.940 align:middle line:84% if you ask, why aren't there any languages like this, 00:10:08.940 --> 00:10:12.930 align:middle line:84% maybe it's just if it's important, 00:10:12.930 --> 00:10:15.353 align:middle line:84% then you want it to be first, something like that. 00:10:15.353 --> 00:10:16.770 align:middle line:84% The theory doesn't usually get any 00:10:16.770 --> 00:10:18.395 align:middle line:90% more sophisticated than that. 00:10:18.395 --> 00:10:20.520 align:middle line:84% So if you're going to do anything with a wh-phrase, 00:10:20.520 --> 00:10:23.340 align:middle line:84% maybe it's not so surprising that it's first. 00:10:23.340 --> 00:10:26.460 align:middle line:84% And so I actually want to show you 00:10:26.460 --> 00:10:28.470 align:middle line:84% a kind of apparent counterexample 00:10:28.470 --> 00:10:30.870 align:middle line:90% to what I've just claimed. 00:10:30.870 --> 00:10:33.990 align:middle line:84% Because the fact is that there are languages 00:10:33.990 --> 00:10:37.410 align:middle line:84% that arrange for the wh-word to be at the end of the sentence, 00:10:37.410 --> 00:10:41.380 align:middle line:84% it's just that they are always wh-in-situ languages. 00:10:41.380 --> 00:10:44.890 align:middle line:84% Let me explain to you what I mean by that. 00:10:44.890 --> 00:10:46.930 align:middle line:84% So in order to tell you what I mean by that, 00:10:46.930 --> 00:10:48.513 align:middle line:84% I first have to tell you another point 00:10:48.513 --> 00:10:50.820 align:middle line:90% of variation between languages. 00:10:50.820 --> 00:10:58.440 align:middle line:84% There are languages in which the normal way to form wh-questions 00:10:58.440 --> 00:11:02.430 align:middle line:84% is it involves what's sometimes called a cleft. 00:11:02.430 --> 00:11:05.280 align:middle line:84% It's as though you have to say in these languages 00:11:05.280 --> 00:11:07.380 align:middle line:84% "What is the one that you bought?" 00:11:07.380 --> 00:11:10.170 align:middle line:84% or What is the thing that you bought?" 00:11:10.170 --> 00:11:11.850 align:middle line:84% Some of the languages that are like this 00:11:11.850 --> 00:11:14.350 align:middle line:84% only do this for some of their kinds of wh-questions. 00:11:14.350 --> 00:11:17.110 align:middle line:84% Some of them do it for all other kinds of wh-questions. 00:11:17.110 --> 00:11:19.380 align:middle line:84% So these are languages in which you can't literally 00:11:19.380 --> 00:11:20.640 align:middle line:90% say "What did you buy?" 00:11:20.640 --> 00:11:23.160 align:middle line:84% You have to do this more complicated structure. 00:11:23.160 --> 00:11:25.740 align:middle line:84% I'm not going to try to give you a tree for this more 00:11:25.740 --> 00:11:26.670 align:middle line:90% complicated structure. 00:11:26.670 --> 00:11:27.907 align:middle line:90% But something like this. 00:11:27.907 --> 00:11:30.240 align:middle line:84% Something that involves more than one clause, basically. 00:11:30.240 --> 00:11:33.450 align:middle line:84% So you have a "what" at the beginning of the sentence. 00:11:33.450 --> 00:11:34.830 align:middle line:84% But inside that sentence, there's 00:11:34.830 --> 00:11:36.630 align:middle line:84% this other clause "that you bought." 00:11:36.630 --> 00:11:37.410 align:middle line:90% That's a CP. 00:11:37.410 --> 00:11:41.730 align:middle line:84% Look, there's a C at the beginning of it-- that. 00:11:41.730 --> 00:11:43.650 align:middle line:84% In English, we can ask questions like, "What 00:11:43.650 --> 00:11:44.817 align:middle line:90% is the one that you bought?" 00:11:44.817 --> 00:11:46.980 align:middle line:84% But in Tagalog, for example, you have to. 00:11:46.980 --> 00:11:49.230 align:middle line:84% So you can't just say, "What did you buy?" in Tagalog. 00:11:49.230 --> 00:11:52.920 align:middle line:84% You have to say, [TAGALOG] where the [TAGALOG] is literally 00:11:52.920 --> 00:11:55.800 align:middle line:84% something like "the (____) you bought." 00:11:55.800 --> 00:11:57.570 align:middle line:84% So that [TAGALOG] is the kind of thing 00:11:57.570 --> 00:12:01.000 align:middle line:84% that only ever goes before noun phrases. 00:12:01.000 --> 00:12:05.310 align:middle line:84% So this is the Tagalog for "What was the thing that you bought?" 00:12:05.310 --> 00:12:06.570 align:middle line:90% or "the one that you bought." 00:12:06.570 --> 00:12:10.260 align:middle line:84% And Tagalog, the word for "one," the word for "thing" is null. 00:12:10.260 --> 00:12:11.310 align:middle line:90% It's not pronounced. 00:12:11.310 --> 00:12:12.570 align:middle line:90% I think I got asked last time. 00:12:12.570 --> 00:12:14.940 align:middle line:84% Are there ever things in trees that are not pronounced. 00:12:14.940 --> 00:12:16.390 align:middle line:90% And here's an example. 00:12:16.390 --> 00:12:18.303 align:middle line:84% The Tagalog word for "one" is such a thing. 00:12:18.303 --> 00:12:19.095 align:middle line:90% AUDIENCE: Question. 00:12:19.095 --> 00:12:21.358 align:middle line:84% So there are [? a ?] type of sentences 00:12:21.358 --> 00:12:23.400 align:middle line:84% which is more like, if you didn't hear something, 00:12:23.400 --> 00:12:24.160 align:middle line:90% you want to clarify. 00:12:24.160 --> 00:12:26.310 align:middle line:84% For example, you want to say the question of, what? 00:12:26.310 --> 00:12:27.210 align:middle line:90% NORVIN RICHARDS: Yes. 00:12:27.210 --> 00:12:27.710 align:middle line:90% Yeah. 00:12:27.710 --> 00:12:31.210 align:middle line:84% AUDIENCE: And I see that there's slightly different structure. 00:12:31.210 --> 00:12:34.000 align:middle line:84% It's very [INAUDIBLE] to say, [INAUDIBLE] or something. 00:12:34.000 --> 00:12:35.250 align:middle line:90% NORVIN RICHARDS: Yeah. 00:12:35.250 --> 00:12:38.685 align:middle line:84% AUDIENCE: But how does that fit with the fact that wh 00:12:38.685 --> 00:12:39.600 align:middle line:90% [INAUDIBLE]. 00:12:39.600 --> 00:12:41.225 align:middle line:84% NORVIN RICHARDS: Is never to the right? 00:12:41.225 --> 00:12:43.900 align:middle line:84% Yeah, so that's a really nice example. 00:12:43.900 --> 00:12:46.380 align:middle line:84% The fact about questions like that 00:12:46.380 --> 00:12:49.950 align:middle line:84% is that the wh-word is sitting where it would normally be. 00:12:49.950 --> 00:12:52.233 align:middle line:84% It's not moving at all, because it 00:12:52.233 --> 00:12:54.150 align:middle line:84% doesn't have to be at the end of the sentence. 00:12:54.150 --> 00:13:01.560 align:middle line:84% If I say, "Mary bought an Estonian novel in Paris," 00:13:01.560 --> 00:13:05.850 align:middle line:84% your response to that could be, "Mary bought what in Paris?" 00:13:05.850 --> 00:13:07.290 align:middle line:84% So the word for "what" isn't going 00:13:07.290 --> 00:13:08.770 align:middle line:90% at the end of the sentence. 00:13:08.770 --> 00:13:09.960 align:middle line:90% In fact, it can't. 00:13:09.960 --> 00:13:13.050 align:middle line:84% It's not the natural way to-- that's the natural way 00:13:13.050 --> 00:13:13.920 align:middle line:90% to respond to that. 00:13:13.920 --> 00:13:17.400 align:middle line:84% You don't say, "Mary bought in Paris what?" 00:13:17.400 --> 00:13:19.470 align:middle line:84% So the word "what" isn't going at the end of-- 00:13:19.470 --> 00:13:22.080 align:middle line:84% it happens to be at the end of the sentence in your example, 00:13:22.080 --> 00:13:25.000 align:middle line:84% but what it's really doing is just not moving. 00:13:25.000 --> 00:13:26.280 align:middle line:90% It's just staying where it is. 00:13:26.280 --> 00:13:27.655 align:middle line:84% That's a really nice point, which 00:13:27.655 --> 00:13:30.300 align:middle line:84% I was planning to make later, but thank you for making it 00:13:30.300 --> 00:13:32.970 align:middle line:84% now, that we have to draw this distinction between just 00:13:32.970 --> 00:13:35.400 align:middle line:84% sitting there and moving to the right. 00:13:35.400 --> 00:13:37.200 align:middle line:84% Yeah, so in those kinds of questions, 00:13:37.200 --> 00:13:38.408 align:middle line:90% it's not moving to the right. 00:13:38.408 --> 00:13:39.360 align:middle line:90% Nice, nice example. 00:13:39.360 --> 00:13:40.710 align:middle line:90% Does that answer your question? 00:13:40.710 --> 00:13:41.580 align:middle line:90% Yeah. 00:13:41.580 --> 00:13:44.100 align:middle line:90% Yeah, no questions? 00:13:44.100 --> 00:13:45.810 align:middle line:90% OK. 00:13:45.810 --> 00:13:48.390 align:middle line:84% OK, so this is the setup for this point. 00:13:48.390 --> 00:13:50.970 align:middle line:84% There are no languages that move wh-words to the right, 00:13:50.970 --> 00:13:54.630 align:middle line:84% but I'm going to show you a weird wiggle that 00:13:54.630 --> 00:13:56.260 align:middle line:90% obscures that fact sometimes. 00:13:56.260 --> 00:13:58.980 align:middle line:84% So first, there are languages like Tagalog 00:13:58.980 --> 00:14:01.608 align:middle line:84% in which you can't ask, "What did you buy?" 00:14:01.608 --> 00:14:03.150 align:middle line:84% You have to say something like, "What 00:14:03.150 --> 00:14:05.250 align:middle line:90% was the thing that you bought?" 00:14:05.250 --> 00:14:07.440 align:middle line:84% And I just have to ask you to trust me 00:14:07.440 --> 00:14:10.650 align:middle line:84% that that's the best analysis of this Tagalog question. 00:14:10.650 --> 00:14:11.430 align:middle line:90% I work on Tagalog. 00:14:11.430 --> 00:14:13.715 align:middle line:84% I'm happy to talk about this at great length. 00:14:13.715 --> 00:14:16.090 align:middle line:84% Talking about Tagalog is one of my favorite things to do. 00:14:16.090 --> 00:14:18.763 align:middle line:84% So don't ask me any questions about this Tagalog question. 00:14:18.763 --> 00:14:19.680 align:middle line:90% That's very dangerous. 00:14:19.680 --> 00:14:22.050 align:middle line:84% If you want to learn anything else today, 00:14:22.050 --> 00:14:25.560 align:middle line:84% we have to get away from this slide, OK? 00:14:25.560 --> 00:14:31.750 align:middle line:84% All right, now, imagine-- so Tagalog is a language that 00:14:31.750 --> 00:14:32.710 align:middle line:90% has these clefts. 00:14:32.710 --> 00:14:36.790 align:middle line:84% It has these bi-clausal ways of asking questions, wh-questions. 00:14:36.790 --> 00:14:38.950 align:middle line:84% And it has obligatory wh-movement. 00:14:38.950 --> 00:14:42.320 align:middle line:84% So the Tagalog wh-words, they're just like the English wh-words. 00:14:42.320 --> 00:14:45.910 align:middle line:84% They have to go at the beginning of the sentence. 00:14:45.910 --> 00:14:50.080 align:middle line:84% Yeah, now imagine what a language would be like, though, 00:14:50.080 --> 00:14:53.380 align:middle line:84% if it had clefting and it had wh-in-situ. 00:14:53.380 --> 00:14:58.130 align:middle line:84% Well, then, it would have questions like this. 00:14:58.130 --> 00:15:01.120 align:middle line:84% It would say things like, "The one that you bought was what?" 00:15:01.120 --> 00:15:03.820 align:middle line:84% Or, "The one that ate the meat was who? 00:15:03.820 --> 00:15:07.520 align:middle line:84% That is, you'd have that embedded clause, 00:15:07.520 --> 00:15:10.630 align:middle line:84% the embedded clause of the cleft, and then the wh-word-- 00:15:10.630 --> 00:15:12.940 align:middle line:84% which is kind of the predicate, the thing 00:15:12.940 --> 00:15:14.980 align:middle line:84% that's being described as the thing 00:15:14.980 --> 00:15:17.530 align:middle line:84% that this embedded clause applies to-- 00:15:17.530 --> 00:15:19.342 align:middle line:84% would end up at the end of the sentence-- 00:15:19.342 --> 00:15:21.550 align:middle line:84% or at least it could, if that's where predicates went 00:15:21.550 --> 00:15:22.758 align:middle line:90% in the language in question-- 00:15:22.758 --> 00:15:26.350 align:middle line:90% 00:15:26.350 --> 00:15:28.270 align:middle line:84% but not because the wh moved there, 00:15:28.270 --> 00:15:29.500 align:middle line:90% just because it was in situ. 00:15:29.500 --> 00:15:32.980 align:middle line:84% It would be like Vlada's example of a second ago, right? 00:15:32.980 --> 00:15:36.490 align:middle line:84% So Vlada was saying, here's an example 00:15:36.490 --> 00:15:38.740 align:middle line:84% where the wh-word is at the end of a sentence. 00:15:38.740 --> 00:15:41.080 align:middle line:84% And I was saying, yes, but it's by accident. 00:15:41.080 --> 00:15:42.820 align:middle line:84% It's really just staying in situ. 00:15:42.820 --> 00:15:45.050 align:middle line:84% It's not moving to the end of the sentence. 00:15:45.050 --> 00:15:47.412 align:middle line:90% Does that make sense? 00:15:47.412 --> 00:15:49.870 align:middle line:84% I'm saying all this because there are languages like this-- 00:15:49.870 --> 00:15:53.050 align:middle line:90% 00:15:53.050 --> 00:15:55.870 align:middle line:84% languages where the wh-word is at the end of the sentence, 00:15:55.870 --> 00:15:57.880 align:middle line:84% not because it's move to the right, 00:15:57.880 --> 00:16:02.922 align:middle line:84% but because, well, the language has clefts, and the wh-word 00:16:02.922 --> 00:16:04.630 align:middle line:84% always ends up at the end of the sentence 00:16:04.630 --> 00:16:07.510 align:middle line:84% not because it moves there, but because it's in situ. 00:16:07.510 --> 00:16:09.530 align:middle line:84% Kabardian is a language like this. 00:16:09.530 --> 00:16:12.100 align:middle line:84% And if you thought I was going to pronounce this Kabardian 00:16:12.100 --> 00:16:14.200 align:middle line:84% example for you, you have another think coming. 00:16:14.200 --> 00:16:16.643 align:middle line:84% I have no hope of pronouncing this Kabardian example. 00:16:16.643 --> 00:16:19.060 align:middle line:84% But this is how you say "Who ate the meat?" in Kabardian-- 00:16:19.060 --> 00:16:22.075 align:middle line:84% the only thing I know how to say in Kabardian where I don't 00:16:22.075 --> 00:16:23.200 align:middle line:90% in fact know how to say it. 00:16:23.200 --> 00:16:25.150 align:middle line:90% I only know what it looks like. 00:16:25.150 --> 00:16:27.790 align:middle line:90% I can't pronounce it. 00:16:27.790 --> 00:16:31.810 align:middle line:84% And it is literally, "The one who ate the meat is who?" 00:16:31.810 --> 00:16:34.480 align:middle line:84% Like that's how you say that in Kabardian. 00:16:34.480 --> 00:16:37.240 align:middle line:84% So there actually are languages that put their wh-words 00:16:37.240 --> 00:16:38.350 align:middle line:90% at the ends of sentences. 00:16:38.350 --> 00:16:39.460 align:middle line:90% Kabardian is one. 00:16:39.460 --> 00:16:43.270 align:middle line:84% But the structure never involves wh-movement to the right. 00:16:43.270 --> 00:16:46.840 align:middle line:84% It's always clefting with the wh-in-situ. 00:16:46.840 --> 00:16:49.870 align:middle line:84% The conversation with the person in the airplane 00:16:49.870 --> 00:16:51.170 align:middle line:90% is always over by this point. 00:16:51.170 --> 00:16:53.795 align:middle line:84% But if they would listen to me, this is what I would tell them. 00:16:53.795 --> 00:16:59.660 align:middle line:84% It's yeah, you don't want a story that says, wh-words-- 00:16:59.660 --> 00:17:01.690 align:middle line:84% it's natural for wh-words to be at the beginning 00:17:01.690 --> 00:17:02.357 align:middle line:90% of the sentence. 00:17:02.357 --> 00:17:03.968 align:middle line:84% First of all, you then have to tell me 00:17:03.968 --> 00:17:05.260 align:middle line:90% what the heck you mean by that. 00:17:05.260 --> 00:17:09.442 align:middle line:84% But there actually are languages in which the wh-word always 00:17:09.442 --> 00:17:10.900 align:middle line:84% ends up at the end of the sentence. 00:17:10.900 --> 00:17:12.654 align:middle line:84% It's just that it doesn't move there. 00:17:12.654 --> 00:17:14.154 align:middle line:84% It's always in situ when it's there. 00:17:14.154 --> 00:17:15.175 align:middle line:90% Faith? 00:17:15.175 --> 00:17:17.170 align:middle line:84% AUDIENCE: So I'm thinking about in Spanish, 00:17:17.170 --> 00:17:20.243 align:middle line:84% how a direct object pronoun can go before or after a verb. 00:17:20.243 --> 00:17:21.160 align:middle line:90% NORVIN RICHARDS: Yeah. 00:17:21.160 --> 00:17:25.215 align:middle line:84% AUDIENCE: But there's always, like, leftward wh-movement, 00:17:25.215 --> 00:17:27.990 align:middle line:84% because you could say, look, "Lo quiero decir" 00:17:27.990 --> 00:17:30.240 align:middle line:84% or "Quiero decirlo," but it's always going to be "¿Qué 00:17:30.240 --> 00:17:30.865 align:middle line:90% quieres decir?" 00:17:30.865 --> 00:17:32.876 align:middle line:90% Is there a reason for that? 00:17:32.876 --> 00:17:37.330 align:middle line:84% NORVIN RICHARDS: So Spanish is a language 00:17:37.330 --> 00:17:39.280 align:middle line:84% that has been described as having wh-movement. 00:17:39.280 --> 00:17:41.980 align:middle line:84% Usually, there's supposed to be contexts 00:17:41.980 --> 00:17:45.520 align:middle line:84% in which it's OK to say the Spanish equivalent of "and you 00:17:45.520 --> 00:17:47.005 align:middle line:90% bought what?" 00:17:47.005 --> 00:17:49.750 align:middle line:84% where the wh-word ends up at the end of the sentence. 00:17:49.750 --> 00:17:53.110 align:middle line:84% And those are supposed to not be exactly like the English cases 00:17:53.110 --> 00:17:56.230 align:middle line:90% of being astonished. 00:17:56.230 --> 00:17:58.540 align:middle line:84% What you really wanted to know about was the pronoun-- 00:17:58.540 --> 00:18:01.450 align:middle line:84% like why can you say, in Spanish, 00:18:01.450 --> 00:18:07.060 align:middle line:84% either "He wants to say it," or "It he wants to say," Right? 00:18:07.060 --> 00:18:09.910 align:middle line:84% So Spanish has these words-- they're called 00:18:09.910 --> 00:18:11.440 align:middle line:90% "clitics" in the literature. 00:18:11.440 --> 00:18:14.860 align:middle line:84% It's common for pronouns to be like this in a lot of languages 00:18:14.860 --> 00:18:17.830 align:middle line:84% where they have special rules about where 00:18:17.830 --> 00:18:19.510 align:middle line:90% they go in the sentence. 00:18:19.510 --> 00:18:22.840 align:middle line:84% I have only shown you one kind of movement, wh-movement, 00:18:22.840 --> 00:18:25.900 align:middle line:84% but what you're pointing out is that there are other kinds. 00:18:25.900 --> 00:18:28.750 align:middle line:84% So Spanish is a language that more or less has to move its 00:18:28.750 --> 00:18:32.050 align:middle line:84% wh-phrases, but its clitics have special conditions on where 00:18:32.050 --> 00:18:35.590 align:middle line:84% they go that are distinct from its-- 00:18:35.590 --> 00:18:36.190 align:middle line:90% from a-- 00:18:36.190 --> 00:18:39.670 align:middle line:84% It's important for your example that it involves an infinitive, 00:18:39.670 --> 00:18:40.360 align:middle line:90% I think-- 00:18:40.360 --> 00:18:45.910 align:middle line:84% that if you wanted to say, whatever, "he said it," 00:18:45.910 --> 00:18:47.260 align:middle line:90% that you would say-- 00:18:47.260 --> 00:18:48.850 align:middle line:90% what would you say-- "it"?-- 00:18:48.850 --> 00:18:49.810 align:middle line:90% "Lo dijo." yeah. 00:18:49.810 --> 00:18:52.270 align:middle line:90% So you wouldn't say "Dijo lo." 00:18:52.270 --> 00:18:56.770 align:middle line:84% AUDIENCE: Yeah, I think that is a thing in Castilian Spanish 00:18:56.770 --> 00:19:00.150 align:middle line:84% from the 1600s, is the [INAUDIBLE].. 00:19:00.150 --> 00:19:01.900 align:middle line:84% NORVIN RICHARDS: It's a point of variation 00:19:01.900 --> 00:19:03.200 align:middle line:90% between Romance languages. 00:19:03.200 --> 00:19:06.175 align:middle line:84% So there are Romance languages-- and languages, generally. 00:19:06.175 --> 00:19:09.340 align:middle line:84% There are Romance-- so these are languages 00:19:09.340 --> 00:19:11.937 align:middle line:84% in which the clitic wants to attach to the verb, 00:19:11.937 --> 00:19:14.020 align:middle line:84% and whether it goes before or after the verb often 00:19:14.020 --> 00:19:17.390 align:middle line:84% depends on whether the verb is an infinitival verb or not. 00:19:17.390 --> 00:19:19.660 align:middle line:84% And for Spanish, I think, the rules 00:19:19.660 --> 00:19:23.800 align:middle line:84% are it goes before tensed verbs and after infinitival verbs. 00:19:23.800 --> 00:19:27.580 align:middle line:84% So your first example had both an infinitive and a tensed verb 00:19:27.580 --> 00:19:29.650 align:middle line:84% in it, and the clitic had two options 00:19:29.650 --> 00:19:31.482 align:middle line:90% as to where it could go. 00:19:31.482 --> 00:19:33.940 align:middle line:84% If we were developing a complete theory of clitics-- which, 00:19:33.940 --> 00:19:36.047 align:middle line:90% just to be clear, we are not-- 00:19:36.047 --> 00:19:38.380 align:middle line:84% we'd want to have an account of that, figuring that out. 00:19:38.380 --> 00:19:40.660 align:middle line:84% The Phenomenon is called "clitic climbing." 00:19:40.660 --> 00:19:42.640 align:middle line:84% The clitic that belongs to the embedded verb 00:19:42.640 --> 00:19:45.382 align:middle line:84% gets to climb up and attach to the first verb. 00:19:45.382 --> 00:19:47.470 align:middle line:90% Was there another hand? 00:19:47.470 --> 00:19:50.070 align:middle line:84% I thought all of you had hands, but I guess 00:19:50.070 --> 00:19:51.940 align:middle line:90% they're all sitting peacefully. 00:19:51.940 --> 00:19:55.860 align:middle line:90% OK, cool. 00:19:55.860 --> 00:19:59.220 align:middle line:84% OK, so language universal, there is no true wh-movement 00:19:59.220 --> 00:19:59.910 align:middle line:90% to the right. 00:19:59.910 --> 00:20:02.100 align:middle line:84% There is Kabardian, so wh-in-situ 00:20:02.100 --> 00:20:03.690 align:middle line:84% combined with the obligatory clefting, 00:20:03.690 --> 00:20:06.240 align:middle line:84% with the consequence that the wh is at the end of the sentence. 00:20:06.240 --> 00:20:08.130 align:middle line:84% But there's never a wh-movement to the right. 00:20:08.130 --> 00:20:10.110 align:middle line:84% If you're a syntactician, you look 00:20:10.110 --> 00:20:13.140 align:middle line:84% at this fact about the world, and you think, well, why? 00:20:13.140 --> 00:20:17.490 align:middle line:84% Why is wh-movement, why is the specifier of CP the-- 00:20:17.490 --> 00:20:22.860 align:middle line:84% we go back to the tree over here, this tree-- 00:20:22.860 --> 00:20:26.010 align:middle line:84% that wh-movement that lands in the specifier of CP 00:20:26.010 --> 00:20:28.770 align:middle line:84% that makes the noun phrase, what, a daughter of CP? 00:20:28.770 --> 00:20:31.770 align:middle line:84% Why does that daughter of CP always precede C-bar? 00:20:31.770 --> 00:20:33.000 align:middle line:90% Why can't it follow C-bar? 00:20:33.000 --> 00:20:35.542 align:middle line:84% There's no language anywhere on Earth where it follows C-bar, 00:20:35.542 --> 00:20:36.990 align:middle line:84% and it would be nice to know why. 00:20:36.990 --> 00:20:39.450 align:middle line:84% So there's syntactic work on why, which 00:20:39.450 --> 00:20:40.590 align:middle line:90% I will not try to show you. 00:20:40.590 --> 00:20:43.200 align:middle line:84% If you would like to know more, take more linguistics classes. 00:20:43.200 --> 00:20:47.150 align:middle line:90% 00:20:47.150 --> 00:20:49.020 align:middle line:90% OK, all right. 00:20:49.020 --> 00:20:52.235 align:middle line:84% And then, yep-- and then, and this 00:20:52.235 --> 00:20:54.360 align:middle line:84% is the part that I went through too fast last time. 00:20:54.360 --> 00:20:55.850 align:middle line:84% And so I'm going to go through it a little more slowly 00:20:55.850 --> 00:20:56.270 align:middle line:90% this time. 00:20:56.270 --> 00:20:56.640 align:middle line:90% I don't know. 00:20:56.640 --> 00:20:57.140 align:middle line:90% We'll see. 00:20:57.140 --> 00:21:00.520 align:middle line:90% 00:21:00.520 --> 00:21:04.840 align:middle line:84% That kind of generalization about languages 00:21:04.840 --> 00:21:09.640 align:middle line:84% not being completely varied, it's not the case. 00:21:09.640 --> 00:21:15.695 align:middle line:84% There-- there's a phenomenon, wh-movement, and there are-- 00:21:15.695 --> 00:21:17.320 align:middle line:84% there's more than one kind of language. 00:21:17.320 --> 00:21:19.923 align:middle line:84% There are languages that do wh-movement like English, 00:21:19.923 --> 00:21:21.340 align:middle line:84% and there are languages that leave 00:21:21.340 --> 00:21:26.930 align:middle line:84% the wh-in-situ, like Mandarin or Kabardian, but that's it. 00:21:26.930 --> 00:21:28.840 align:middle line:84% Those are all the kinds that there are. 00:21:28.840 --> 00:21:30.930 align:middle line:84% And there are imaginable kinds, like the mirror 00:21:30.930 --> 00:21:32.680 align:middle line:84% image of English, where the wh-phrase goes 00:21:32.680 --> 00:21:34.540 align:middle line:90% at the end, which don't exist. 00:21:34.540 --> 00:21:36.033 align:middle line:90% And we'd like to know why. 00:21:36.033 --> 00:21:38.200 align:middle line:84% I want to show you another example of the same kind. 00:21:38.200 --> 00:21:41.260 align:middle line:84% I showed you this last time, but I'll show it to you again. 00:21:41.260 --> 00:21:42.880 align:middle line:84% There's also variation with respect 00:21:42.880 --> 00:21:45.110 align:middle line:84% to how you do multiple wh-questions, 00:21:45.110 --> 00:21:47.050 align:middle line:84% So wh-questions where there's more than one 00:21:47.050 --> 00:21:49.762 align:middle line:90% wh-word in the sentence. 00:21:49.762 --> 00:21:51.970 align:middle line:84% First of all, there are languages that just ban them. 00:21:51.970 --> 00:21:53.480 align:middle line:90% That's one kind of language. 00:21:53.480 --> 00:21:56.140 align:middle line:90% So Irish doesn't allow them. 00:21:56.140 --> 00:21:58.330 align:middle line:90% Italian, at least until-- 00:21:58.330 --> 00:22:00.070 align:middle line:84% apparently, young Italians are beginning 00:22:00.070 --> 00:22:01.990 align:middle line:84% to develop multiple-wh questions. 00:22:01.990 --> 00:22:03.370 align:middle line:90% [LAUGHTER] 00:22:03.370 --> 00:22:06.340 align:middle line:84% But older Italians don't have them. 00:22:06.340 --> 00:22:09.760 align:middle line:84% Kids today and their multiple-wh questions! 00:22:09.760 --> 00:22:12.970 align:middle line:84% And then, so for languages that do have them, 00:22:12.970 --> 00:22:16.720 align:middle line:84% they're languages like English in which the rule is, 00:22:16.720 --> 00:22:19.420 align:middle line:84% English has wh-movement, as we know, 00:22:19.420 --> 00:22:23.222 align:middle line:84% and if you have more than one wh-word, the rule for English 00:22:23.222 --> 00:22:24.430 align:middle line:90% is that you move one of them. 00:22:24.430 --> 00:22:28.030 align:middle line:84% So you ask questions like, "What did you give to whom?" 00:22:28.030 --> 00:22:30.610 align:middle line:84% Or "What did you give to whom when?" 00:22:30.610 --> 00:22:32.288 align:middle line:84% If you have three wh-phrases, you're 00:22:32.288 --> 00:22:33.580 align:middle line:90% only going to move one of them. 00:22:33.580 --> 00:22:36.660 align:middle line:84% That's the generalization about English. 00:22:36.660 --> 00:22:38.700 align:middle line:84% And then-- I told you this last time-- 00:22:38.700 --> 00:22:41.310 align:middle line:84% there are languages like Bulgarian and Mohawk 00:22:41.310 --> 00:22:43.470 align:middle line:84% in which all of the wh-words move. 00:22:43.470 --> 00:22:47.040 align:middle line:84% So in Bulgarian, you must say, "What to whom did he give?" 00:22:47.040 --> 00:22:48.420 align:middle line:90% That's that second question. 00:22:48.420 --> 00:22:50.040 align:middle line:84% Or in Mohawk, you say things like, 00:22:50.040 --> 00:22:52.200 align:middle line:90% "Tell me who what bought." 00:22:52.200 --> 00:22:55.005 align:middle line:84% You can't leave either of the wh-phrases in situ. 00:22:55.005 --> 00:22:58.210 align:middle line:90% 00:22:58.210 --> 00:23:00.850 align:middle line:84% And so here's the place where I promised you, 00:23:00.850 --> 00:23:04.010 align:middle line:84% there's more than one kind of language, 00:23:04.010 --> 00:23:06.352 align:middle line:84% but there isn't every imaginable kind of language. 00:23:06.352 --> 00:23:07.810 align:middle line:84% So there are languages like English 00:23:07.810 --> 00:23:09.040 align:middle line:90% that move one wh-phrase. 00:23:09.040 --> 00:23:11.290 align:middle line:84% There are languages like Mandarin that don't move any. 00:23:11.290 --> 00:23:13.780 align:middle line:84% And there are languages like Bulgarian and Mohawk 00:23:13.780 --> 00:23:15.040 align:middle line:90% that move all of them. 00:23:15.040 --> 00:23:20.260 align:middle line:84% But 0 and 1 and "all" are all of the options 00:23:20.260 --> 00:23:23.470 align:middle line:84% for moving wh-phrases in multiple wh-questions. 00:23:23.470 --> 00:23:27.070 align:middle line:84% And you all know more math than I do. 00:23:27.070 --> 00:23:29.440 align:middle line:84% You know of numbers besides 0 and 1 00:23:29.440 --> 00:23:32.800 align:middle line:84% and, well, "all," which I guess isn't a number. 00:23:32.800 --> 00:23:34.630 align:middle line:84% So you could imagine a language where 00:23:34.630 --> 00:23:38.230 align:middle line:84% the rule is, move any number of wh-phrases 00:23:38.230 --> 00:23:40.300 align:middle line:90% up to a maximum of 2, right? 00:23:40.300 --> 00:23:43.210 align:middle line:84% Or move any number of wh-phrases, I don't know, 00:23:43.210 --> 00:23:46.960 align:middle line:84% as long as it's prime, or whatever-- 00:23:46.960 --> 00:23:50.080 align:middle line:84% as long as it's one of the first 10 digits of pi. 00:23:50.080 --> 00:23:52.810 align:middle line:84% All kinds of imaginable languages 00:23:52.810 --> 00:23:56.290 align:middle line:84% that you could imagine, and none of them exist, right? 00:23:56.290 --> 00:23:58.090 align:middle line:84% So there aren't languages out there 00:23:58.090 --> 00:24:01.330 align:middle line:84% in which you say "Who gave a book to Mary?" 00:24:01.330 --> 00:24:03.760 align:middle line:84% and you say "Who what gave to Mary?" 00:24:03.760 --> 00:24:06.310 align:middle line:84% but when you get to three, you have 00:24:06.310 --> 00:24:08.050 align:middle line:90% to say "Who what gave to whom?" 00:24:08.050 --> 00:24:09.310 align:middle line:90% That language doesn't exist. 00:24:09.310 --> 00:24:10.360 align:middle line:90% It's not hard to imagine. 00:24:10.360 --> 00:24:11.735 align:middle line:84% You could write a grammar for it. 00:24:11.735 --> 00:24:12.760 align:middle line:90% It doesn't exist. 00:24:12.760 --> 00:24:15.980 align:middle line:90% You don't find it. 00:24:15.980 --> 00:24:18.370 align:middle line:84% And then, this is where I gave you 00:24:18.370 --> 00:24:21.010 align:middle line:84% the parable of the function, which 00:24:21.010 --> 00:24:24.640 align:middle line:84% was meant to be a parable about what life 00:24:24.640 --> 00:24:26.800 align:middle line:90% is like when you are a baby. 00:24:26.800 --> 00:24:30.790 align:middle line:84% So imagine, I said, that you are a Bulgarian baby, 00:24:30.790 --> 00:24:35.410 align:middle line:84% and you're hearing your parents utter multiple wh-questions. 00:24:35.410 --> 00:24:39.370 align:middle line:84% There's some limit on the amount of data that you get-- 00:24:39.370 --> 00:24:41.710 align:middle line:84% not just if you're a Bulgarian baby 00:24:41.710 --> 00:24:43.210 align:middle line:84% listening to multiple wh-questions, 00:24:43.210 --> 00:24:45.190 align:middle line:90% but in life, right? 00:24:45.190 --> 00:24:47.660 align:middle line:84% There is some-- it may not have seemed like it at the time, 00:24:47.660 --> 00:24:50.350 align:middle line:84% but there is some maximum amount of stuff that your parents said 00:24:50.350 --> 00:24:54.048 align:middle line:84% to you when you were growing up, and the stuff that you heard-- 00:24:54.048 --> 00:24:56.590 align:middle line:84% not just your parents, but the other people that were talking 00:24:56.590 --> 00:24:58.120 align:middle line:84% to you when you were growing up and acquiring 00:24:58.120 --> 00:24:59.080 align:middle line:90% your native language. 00:24:59.080 --> 00:25:01.810 align:middle line:84% There is some body of data-- maybe it was fairly large-- 00:25:01.810 --> 00:25:03.880 align:middle line:84% of things that people said to you. 00:25:03.880 --> 00:25:06.280 align:middle line:84% Maybe fairly large, but it was finite, 00:25:06.280 --> 00:25:10.900 align:middle line:84% so there was some maximum number of wh-words in the questions 00:25:10.900 --> 00:25:13.180 align:middle line:90% that you heard growing up. 00:25:13.180 --> 00:25:15.910 align:middle line:84% And the thing about finite data-- 00:25:15.910 --> 00:25:18.640 align:middle line:84% so the point of the parable of the function-- is here, 00:25:18.640 --> 00:25:19.990 align:middle line:90% I am giving you finite data. 00:25:19.990 --> 00:25:22.930 align:middle line:84% The function returns 1 for 1, and 2 for 2, and 3 for 3, 00:25:22.930 --> 00:25:25.630 align:middle line:84% and 4 for 4, and you have no idea at all 00:25:25.630 --> 00:25:28.510 align:middle line:90% what it should return for 5. 00:25:28.510 --> 00:25:31.150 align:middle line:84% Given what I've given you so far, 00:25:31.150 --> 00:25:34.030 align:middle line:84% you might hope that you're living in just 00:25:34.030 --> 00:25:36.970 align:middle line:84% and merciful universe where people, data work 00:25:36.970 --> 00:25:38.000 align:middle line:90% the way they should. 00:25:38.000 --> 00:25:42.460 align:middle line:84% But they're just-- as a matter of logic or a matter of math, 00:25:42.460 --> 00:25:45.150 align:middle line:84% there are no restrictions at all on what 00:25:45.150 --> 00:25:49.420 align:middle line:84% the consequence of giving this function 5 could be. 00:25:49.420 --> 00:25:51.790 align:middle line:84% So I showed you one example, one function 00:25:51.790 --> 00:25:54.400 align:middle line:84% for which the output would be 29, 00:25:54.400 --> 00:25:57.670 align:middle line:84% and I could multiply that first thing, 00:25:57.670 --> 00:25:59.860 align:middle line:84% the thing that comes right after the equals sign, 00:25:59.860 --> 00:26:02.260 align:middle line:84% I could multiply that by any number at all. 00:26:02.260 --> 00:26:05.860 align:middle line:84% The consequence of f of 5 could just be anything. 00:26:05.860 --> 00:26:07.690 align:middle line:90% And life is kind of like that. 00:26:07.690 --> 00:26:10.660 align:middle line:84% This is sometimes called the problem of induction 00:26:10.660 --> 00:26:13.600 align:middle line:84% when people talk about it in the context of science. 00:26:13.600 --> 00:26:18.730 align:middle line:84% So you're going to make some number of observations 00:26:18.730 --> 00:26:22.390 align:middle line:84% of the world, and you'll maybe see the same thing 00:26:22.390 --> 00:26:23.290 align:middle line:90% over and over again. 00:26:23.290 --> 00:26:28.750 align:middle line:84% You'll see that, as you drop a weight, 00:26:28.750 --> 00:26:31.750 align:middle line:84% that the weight falls down, and not sideways or up. 00:26:31.750 --> 00:26:33.790 align:middle line:84% And you see that over, and over, and over again. 00:26:33.790 --> 00:26:36.250 align:middle line:84% And eventually, you start to hypothesize 00:26:36.250 --> 00:26:39.850 align:middle line:84% that down is the direction that they're going to fall, 00:26:39.850 --> 00:26:43.180 align:middle line:84% that every crow you see is going to be black, 00:26:43.180 --> 00:26:45.070 align:middle line:84% that every swan you see is going to be white, 00:26:45.070 --> 00:26:47.500 align:middle line:84% because you keep seeing that over and over and over again. 00:26:47.500 --> 00:26:51.850 align:middle line:84% But it's kind of like concluding that f of 5 is going to be 5. 00:26:51.850 --> 00:26:54.910 align:middle line:84% It should work that way given everything that you've seen, 00:26:54.910 --> 00:26:56.200 align:middle line:90% but it might not. 00:26:56.200 --> 00:27:00.280 align:middle line:84% Maybe the next weight you drop will fall upwards. 00:27:00.280 --> 00:27:03.100 align:middle line:84% And so part of science is saying, 00:27:03.100 --> 00:27:04.960 align:middle line:90% we're going to draw these laws. 00:27:04.960 --> 00:27:08.680 align:middle line:84% Given the inevitably finite data that we've got, 00:27:08.680 --> 00:27:10.963 align:middle line:84% we're going to sketch these laws about what 00:27:10.963 --> 00:27:12.130 align:middle line:90% we think is going to happen. 00:27:12.130 --> 00:27:16.840 align:middle line:84% And then we go on doing science and refining our laws. 00:27:16.840 --> 00:27:18.040 align:middle line:90% Raquel, are you-- yeah? 00:27:18.040 --> 00:27:19.720 align:middle line:84% AUDIENCE: The thought that I am thinking 00:27:19.720 --> 00:27:21.730 align:middle line:84% is something along the lines of like, 00:27:21.730 --> 00:27:23.980 align:middle line:84% is it possible that part of being human 00:27:23.980 --> 00:27:28.450 align:middle line:84% is developing languages that are kinder than giving you "f of 5 00:27:28.450 --> 00:27:34.570 align:middle line:84% is 29," and that maybe it's like you could give languages 00:27:34.570 --> 00:27:39.550 align:middle line:84% a weird difficulty coefficient, and then the coefficient, 00:27:39.550 --> 00:27:43.100 align:middle line:84% if it's really low, tells you that it'll probably be 00:27:43.100 --> 00:27:44.350 align:middle line:90% something intuitive like five. 00:27:44.350 --> 00:27:46.420 align:middle line:84% If it's really high, then you can't 00:27:46.420 --> 00:27:49.150 align:middle line:84% rely on the language to be nice to you when it could be 29? 00:27:49.150 --> 00:27:50.980 align:middle line:84% NORVIN RICHARDS: Yeah, that's-- you're saying it very well. 00:27:50.980 --> 00:27:53.740 align:middle line:84% Maybe another way to say it, to repeat what you just said, 00:27:53.740 --> 00:27:58.900 align:middle line:84% is yeah, human minds are set up in such a way that f of 5 is 5. 00:27:58.900 --> 00:28:02.710 align:middle line:84% Human minds are set up in such a way that multiple wh-questions, 00:28:02.710 --> 00:28:04.930 align:middle line:90% there are only a few options. 00:28:04.930 --> 00:28:07.840 align:middle line:84% Maybe I'm learning Italian and multiple wh-questions 00:28:07.840 --> 00:28:08.710 align:middle line:90% are just banned. 00:28:08.710 --> 00:28:12.130 align:middle line:84% Maybe I'm learning Mandarin and wh-phrases all stay in situ. 00:28:12.130 --> 00:28:14.200 align:middle line:84% Maybe I'm learning English and one of them moves. 00:28:14.200 --> 00:28:16.283 align:middle line:84% Maybe I'm learning Bulgarian and all of them move. 00:28:16.283 --> 00:28:17.150 align:middle line:90% And that's it. 00:28:17.150 --> 00:28:19.510 align:middle line:84% Those are the only options there are. 00:28:19.510 --> 00:28:21.533 align:middle line:84% Which means if you're a Bulgarian child, 00:28:21.533 --> 00:28:23.950 align:middle line:84% and you're growing up, and you're hearing your parents ask 00:28:23.950 --> 00:28:27.280 align:middle line:84% multiple wh-questions, there's some maximum number 00:28:27.280 --> 00:28:30.430 align:middle line:84% of wh-words you ever heard them say at any point in their life. 00:28:30.430 --> 00:28:34.930 align:middle line:84% And that number-- suppose that's the most inquisitive 00:28:34.930 --> 00:28:36.520 align:middle line:84% your parents ever got in your life, 00:28:36.520 --> 00:28:39.520 align:middle line:90% It was a two-wh-word question. 00:28:39.520 --> 00:28:43.850 align:middle line:84% I can say, incidentally, multiple-wh questions-- 00:28:43.850 --> 00:28:46.330 align:middle line:84% I wrote my dissertation about multiple-wh questions. 00:28:46.330 --> 00:28:49.570 align:middle line:84% They're extremely interesting, but they're not common. 00:28:49.570 --> 00:28:54.580 align:middle line:84% If you spend some time listening, 00:28:54.580 --> 00:28:56.980 align:middle line:84% you'll spend a long time before you hear 00:28:56.980 --> 00:28:58.720 align:middle line:90% somebody say this in the wild. 00:28:58.720 --> 00:29:02.170 align:middle line:84% It's not a common kind of question to ask. 00:29:02.170 --> 00:29:04.090 align:middle line:84% I happen to know, for example, that there 00:29:04.090 --> 00:29:07.330 align:middle line:84% are no multiple wh-questions in the Bible. 00:29:07.330 --> 00:29:08.575 align:middle line:90% So I have this project-- 00:29:08.575 --> 00:29:09.622 align:middle line:90% [LAUGHTER] 00:29:09.622 --> 00:29:11.580 align:middle line:84% --that involves going through the Bible verses. 00:29:11.580 --> 00:29:13.375 align:middle line:84% So there's a language called Wampanoag, 00:29:13.375 --> 00:29:16.030 align:middle line:84% which is a language that's spoken around here 00:29:16.030 --> 00:29:16.870 align:middle line:90% traditionally. 00:29:16.870 --> 00:29:19.330 align:middle line:84% The Wampanoag are trying to revive their language, and one 00:29:19.330 --> 00:29:22.240 align:middle line:84% of the texts that they have is a complete Bible translation. 00:29:22.240 --> 00:29:24.840 align:middle line:84% It was the first Bible published in this hemisphere. 00:29:24.840 --> 00:29:27.430 align:middle line:84% It was a Wampanoag Bible that was published here 00:29:27.430 --> 00:29:30.550 align:middle line:90% in Boston in the 1600s-- 00:29:30.550 --> 00:29:32.500 align:middle line:90% 1635. 00:29:32.500 --> 00:29:34.720 align:middle line:84% And so I have been reading the Bible very slowly 00:29:34.720 --> 00:29:37.300 align:middle line:84% and carefully, trying to put the grammar of Wampanoag 00:29:37.300 --> 00:29:39.020 align:middle line:84% back together and writing a dictionary. 00:29:39.020 --> 00:29:42.518 align:middle line:84% And so I really wish that there were multiple-wh questions 00:29:42.518 --> 00:29:43.060 align:middle line:90% in the Bible. 00:29:43.060 --> 00:29:45.143 align:middle line:84% It would answer certain questions about Wampanoag. 00:29:45.143 --> 00:29:48.038 align:middle line:84% I would really like Jesus Christ to turn to the disciples 00:29:48.038 --> 00:29:49.330 align:middle line:90% and ask them "Who bought what?" 00:29:49.330 --> 00:29:50.410 align:middle line:90% It would be great. 00:29:50.410 --> 00:29:51.040 align:middle line:90% [LAUGHTER] 00:29:51.040 --> 00:29:52.540 align:middle line:90% But he never does, never. 00:29:52.540 --> 00:29:53.990 align:middle line:90% I have checked. 00:29:53.990 --> 00:29:55.990 align:middle line:84% There are no multiple-wh questions in the Bible. 00:29:55.990 --> 00:29:57.550 align:middle line:84% And the Bible is quite large, right? 00:29:57.550 --> 00:29:58.902 align:middle line:90% There's lots and lots of texts. 00:29:58.902 --> 00:30:00.610 align:middle line:84% I know this because I've been reading it. 00:30:00.610 --> 00:30:01.443 align:middle line:90% It's taken me years. 00:30:01.443 --> 00:30:04.330 align:middle line:90% 00:30:04.330 --> 00:30:05.770 align:middle line:90% So these are not common. 00:30:05.770 --> 00:30:08.590 align:middle line:84% It's not at all-- it wouldn't be at all surprising 00:30:08.590 --> 00:30:11.410 align:middle line:84% if the largest multiple-wh question you ever 00:30:11.410 --> 00:30:13.605 align:middle line:84% heard your parents say had two wh-phrases in it-- 00:30:13.605 --> 00:30:14.980 align:middle line:84% something like "Who bought what?" 00:30:14.980 --> 00:30:16.960 align:middle line:90% or "What did he give to whom?" 00:30:16.960 --> 00:30:18.910 align:middle line:84% Maybe you heard your parents at some point 00:30:18.910 --> 00:30:22.630 align:middle line:84% say, "Who gave what to whom, when, why?" 00:30:22.630 --> 00:30:24.010 align:middle line:90% It's not likely. 00:30:24.010 --> 00:30:26.620 align:middle line:84% And so I think I said this last time. 00:30:26.620 --> 00:30:29.440 align:middle line:84% You might have expected if children 00:30:29.440 --> 00:30:32.200 align:middle line:84% had to consider all of the imaginable grammars that 00:30:32.200 --> 00:30:34.390 align:middle line:84% are consistent with the data that they have, 00:30:34.390 --> 00:30:38.620 align:middle line:84% you might have imagined that Bulgarian children would have-- 00:30:38.620 --> 00:30:41.163 align:middle line:84% Bulgarian adults would have different grammars, 00:30:41.163 --> 00:30:42.580 align:middle line:84% that there would be Bulgarians who 00:30:42.580 --> 00:30:46.330 align:middle line:84% guessed that the grammar was move your wh-phrases up 00:30:46.330 --> 00:30:48.668 align:middle line:84% to a maximum of two, or move your wh-phrases up 00:30:48.668 --> 00:30:50.710 align:middle line:84% to a maximum of three, or move your wh-phrases up 00:30:50.710 --> 00:30:51.585 align:middle line:90% to a maximum of four. 00:30:51.585 --> 00:30:52.990 align:middle line:90% But that's not what we find. 00:30:52.990 --> 00:30:56.410 align:middle line:84% So any Bulgarian that you ask, as you ratchet up 00:30:56.410 --> 00:31:00.400 align:middle line:84% the number of wh-words in the question, they just always 00:31:00.400 --> 00:31:01.930 align:middle line:90% move them all. 00:31:01.930 --> 00:31:03.940 align:middle line:84% The only choice they all made was all. 00:31:03.940 --> 00:31:10.450 align:middle line:84% Yeah, so here's a case where the data that anybody is exposed to 00:31:10.450 --> 00:31:13.540 align:middle line:84% are necessarily limited in a way that 00:31:13.540 --> 00:31:16.660 align:middle line:84% means that there are literally infinitely many possible 00:31:16.660 --> 00:31:19.150 align:middle line:84% grammars that are compatible with the data that they have. 00:31:19.150 --> 00:31:22.300 align:middle line:84% And yet, they all converge on the same answer. 00:31:22.300 --> 00:31:26.170 align:middle line:84% So this is a case of what Raquel was just talking about. 00:31:26.170 --> 00:31:28.083 align:middle line:84% It's as though being a human being 00:31:28.083 --> 00:31:29.500 align:middle line:84% means having the kind of mind that 00:31:29.500 --> 00:31:31.940 align:middle line:84% considers some possibilities but not others. 00:31:31.940 --> 00:31:35.650 align:middle line:84% So you consider all, but you don't consider up 00:31:35.650 --> 00:31:36.850 align:middle line:90% to a maximum of two. 00:31:36.850 --> 00:31:39.660 align:middle line:84% That's not a kind of grammar you think about. 00:31:39.660 --> 00:31:42.040 align:middle line:84% That's apparently what we're finding. 00:31:42.040 --> 00:31:43.870 align:middle line:84% Raising questions like, well, why? 00:31:43.870 --> 00:31:46.630 align:middle line:84% So what is it about the human mind that makes it like that? 00:31:46.630 --> 00:31:51.750 align:middle line:84% That's one of the kinds of things linguists think about. 00:31:51.750 --> 00:31:56.520 align:middle line:84% OK, similarly-- and this is echoing a point 00:31:56.520 --> 00:31:59.500 align:middle line:84% that Vlada made a second ago-- if you're learning Chinese, 00:31:59.500 --> 00:32:01.675 align:middle line:84% if you're growing up learning Chinese, 00:32:01.675 --> 00:32:03.300 align:middle line:84% you may hear your parents say something 00:32:03.300 --> 00:32:05.740 align:middle line:90% like this at some point. 00:32:05.740 --> 00:32:11.010 align:middle line:84% And if you are considering all of the logically possible 00:32:11.010 --> 00:32:12.930 align:middle line:84% grammars that are consistent with this, 00:32:12.930 --> 00:32:15.420 align:middle line:84% you might have to spend some time thinking, well, 00:32:15.420 --> 00:32:17.340 align:middle line:84% am I learning a language that moves 00:32:17.340 --> 00:32:19.860 align:middle line:84% its wh-phrases to the right, or am I 00:32:19.860 --> 00:32:23.160 align:middle line:84% learning a language that leaves its wh-phrases in situ? 00:32:23.160 --> 00:32:24.900 align:middle line:84% And so you might expect that there 00:32:24.900 --> 00:32:29.100 align:middle line:84% would be a stage in the life of Mandarin children 00:32:29.100 --> 00:32:32.250 align:middle line:84% where they entertained the possibility that wh-phrases 00:32:32.250 --> 00:32:33.100 align:middle line:90% move to the right. 00:32:33.100 --> 00:32:36.180 align:middle line:84% It would be a cute kid thing to do. 00:32:36.180 --> 00:32:39.210 align:middle line:90% Oh, our child is so cute. 00:32:39.210 --> 00:32:41.913 align:middle line:84% When she was three, she thought wh-phrases moved to the right. 00:32:41.913 --> 00:32:43.080 align:middle line:90% She was saying things like-- 00:32:43.080 --> 00:32:43.420 align:middle line:90% AUDIENCE: Aww. 00:32:43.420 --> 00:32:44.545 align:middle line:90% NORVIN RICHARDS: Aww, yeah. 00:32:44.545 --> 00:32:45.840 align:middle line:90% [LAUGHTER] 00:32:45.840 --> 00:32:48.690 align:middle line:84% But no, I mean, Chinese babies are cute like babies 00:32:48.690 --> 00:32:51.390 align:middle line:84% everywhere, but they do not make that particular cute mistake. 00:32:51.390 --> 00:32:53.280 align:middle line:84% If you've ever been around small children, 00:32:53.280 --> 00:32:55.440 align:middle line:84% you know children do make mistakes. 00:32:55.440 --> 00:32:59.070 align:middle line:84% They take guesses that are wrong about how the adult 00:32:59.070 --> 00:33:01.920 align:middle line:84% language works, and they spend some time 00:33:01.920 --> 00:33:03.240 align:middle line:90% making adorable mistakes. 00:33:03.240 --> 00:33:05.580 align:middle line:84% But the children do not make that particular adorable 00:33:05.580 --> 00:33:08.910 align:middle line:84% mistake because they are human beings, 00:33:08.910 --> 00:33:11.370 align:middle line:84% and they know that there are on some level-- 00:33:11.370 --> 00:33:14.340 align:middle line:84% their heads are structured in such a way that they cannot 00:33:14.340 --> 00:33:16.770 align:middle line:84% entertain grammars where wh-phrases move to the right. 00:33:16.770 --> 00:33:18.560 align:middle line:90% There aren't any of those. 00:33:18.560 --> 00:33:20.310 align:middle line:84% Part of our job as linguists is to come up 00:33:20.310 --> 00:33:22.540 align:middle line:84% with a theory that explains that. 00:33:22.540 --> 00:33:25.380 align:middle line:84% And when I say it's part of our job as linguists, 00:33:25.380 --> 00:33:29.850 align:middle line:84% I do not mean that I'm now about to tell you the theory. 00:33:29.850 --> 00:33:32.250 align:middle line:84% That's an active topic of research, 00:33:32.250 --> 00:33:35.340 align:middle line:84% people trying to figure out why grammar works that way. 00:33:35.340 --> 00:33:38.220 align:middle line:84% OK so yeah, if you are a Chinese baby 00:33:38.220 --> 00:33:39.960 align:middle line:84% and you hear that sentence, you know 00:33:39.960 --> 00:33:42.420 align:middle line:84% that you are hearing a wh-in-situ language. 00:33:42.420 --> 00:33:44.520 align:middle line:84% You don't have to entertain possibilities 00:33:44.520 --> 00:33:47.370 align:middle line:84% like a language in which the wh-phrase moves to the right, 00:33:47.370 --> 00:33:49.230 align:middle line:84% or, a language in which the wh-phrase is 00:33:49.230 --> 00:33:51.610 align:middle line:90% the third word in the sentence. 00:33:51.610 --> 00:33:55.500 align:middle line:84% Those are both grammars that would generate that question, 00:33:55.500 --> 00:33:57.330 align:middle line:84% but you don't spend any time entertaining 00:33:57.330 --> 00:33:59.100 align:middle line:90% those possibilities. 00:33:59.100 --> 00:33:59.615 align:middle line:90% Yeah? 00:33:59.615 --> 00:34:00.990 align:middle line:84% AUDIENCE: OK, well, this is going 00:34:00.990 --> 00:34:03.483 align:middle line:84% to be probably a bit like what if. 00:34:03.483 --> 00:34:04.400 align:middle line:90% NORVIN RICHARDS: Yeah. 00:34:04.400 --> 00:34:06.100 align:middle line:84% AUDIENCE: Probably never encounter it. 00:34:06.100 --> 00:34:08.239 align:middle line:90% NORVIN RICHARDS: Yeah. 00:34:08.239 --> 00:34:11.139 align:middle line:84% AUDIENCE: Because you'd think that nowhere 00:34:11.139 --> 00:34:17.380 align:middle line:84% does [INAUDIBLE] Chinese babies say, or like, think, 00:34:17.380 --> 00:34:19.715 align:middle line:90% I must move word to end. 00:34:19.715 --> 00:34:21.219 align:middle line:90% NORVIN RICHARDS: Yeah. 00:34:21.219 --> 00:34:26.020 align:middle line:84% AUDIENCE: What if, hypothetically, a baby speaking 00:34:26.020 --> 00:34:27.820 align:middle line:90% in an English-- 00:34:27.820 --> 00:34:32.020 align:middle line:84% a baby in an English-speaking environment never heard 00:34:32.020 --> 00:34:34.320 align:middle line:90% an in-situ-- 00:34:34.320 --> 00:34:35.660 align:middle line:90% NORVIN RICHARDS: wh, mm-hmm. 00:34:35.660 --> 00:34:38.630 align:middle line:90% 00:34:38.630 --> 00:34:41.540 align:middle line:84% AUDIENCE: OK, often-- OK, how do I phrase it. 00:34:41.540 --> 00:34:43.889 align:middle line:90% What if the only-- 00:34:43.889 --> 00:34:46.820 align:middle line:84% all the questions the baby hears is 00:34:46.820 --> 00:34:51.290 align:middle line:84% like, "Who did this," "Who did that," "What happened?" 00:34:51.290 --> 00:34:54.130 align:middle line:84% NORVIN RICHARDS: "What happened," yeah. 00:34:54.130 --> 00:34:57.083 align:middle line:84% AUDIENCE: Where it could plausibly-- 00:34:57.083 --> 00:34:58.250 align:middle line:90% NORVIN RICHARDS: Be in situ. 00:34:58.250 --> 00:34:59.060 align:middle line:90% AUDIENCE: Be in situ. 00:34:59.060 --> 00:34:59.977 align:middle line:90% NORVIN RICHARDS: Yeah. 00:34:59.977 --> 00:35:06.860 align:middle line:84% AUDIENCE: Could they potentially realize, or like think maybe 00:35:06.860 --> 00:35:08.148 align:middle line:90% it's not in situ. 00:35:08.148 --> 00:35:09.440 align:middle line:90% NORVIN RICHARDS: Nice question. 00:35:09.440 --> 00:35:10.980 align:middle line:84% So everything I've told you so far 00:35:10.980 --> 00:35:15.440 align:middle line:84% makes it sound like you could fool babies, in theory, 00:35:15.440 --> 00:35:16.850 align:middle line:90% in that way. 00:35:16.850 --> 00:35:19.870 align:middle line:84% Like you say, it's hard to test because, in fact, I 00:35:19.870 --> 00:35:24.170 align:middle line:84% mean you probably can't recruit parents to agree never 00:35:24.170 --> 00:35:26.360 align:middle line:84% to ask questions like "What do you want for lunch?" 00:35:26.360 --> 00:35:29.060 align:middle line:90% 00:35:29.060 --> 00:35:31.760 align:middle line:84% But you're absolutely right that you could imagine children 00:35:31.760 --> 00:35:34.550 align:middle line:84% going through, probably, a pretty brief stage 00:35:34.550 --> 00:35:39.620 align:middle line:84% where they thought that was a viable option. 00:35:39.620 --> 00:35:44.190 align:middle line:84% In a way, your question gets at something else that I said, 00:35:44.190 --> 00:35:47.030 align:middle line:84% which is languages-- so I've said there are 00:35:47.030 --> 00:35:48.380 align:middle line:90% different kinds of languages. 00:35:48.380 --> 00:35:53.335 align:middle line:84% There are wh-in-situ languages, there are wh-moving languages, 00:35:53.335 --> 00:35:54.710 align:middle line:84% and-- this was Faith's question-- 00:35:54.710 --> 00:35:57.170 align:middle line:84% there are languages that allow either option. 00:35:57.170 --> 00:36:00.880 align:middle line:90% That's also a kind of language. 00:36:00.880 --> 00:36:04.370 align:middle line:84% And that's all I said about that right. 00:36:04.370 --> 00:36:07.598 align:middle line:84% You could now ask, does that just 00:36:07.598 --> 00:36:09.140 align:middle line:84% mean that if I want to learn English, 00:36:09.140 --> 00:36:11.510 align:middle line:84% I must learn it's a wh-movement language? 00:36:11.510 --> 00:36:14.060 align:middle line:84% Or does the fact that it's a wh-movement language 00:36:14.060 --> 00:36:16.377 align:middle line:84% have anything to do with anything else about English? 00:36:16.377 --> 00:36:18.710 align:middle line:84% Is that just a parameter we have to state about English, 00:36:18.710 --> 00:36:21.770 align:middle line:84% or does it follow from other stuff about English? 00:36:21.770 --> 00:36:24.248 align:middle line:84% And that's also an active research question, 00:36:24.248 --> 00:36:26.540 align:middle line:84% is whether we can get away without having to just state 00:36:26.540 --> 00:36:30.102 align:middle line:84% those parameters, or whether we can connect that 00:36:30.102 --> 00:36:31.310 align:middle line:90% to other facts about English. 00:36:31.310 --> 00:36:32.960 align:middle line:84% That would mean that an English-speaking baby would 00:36:32.960 --> 00:36:34.377 align:middle line:84% have other sources of information, 00:36:34.377 --> 00:36:38.450 align:middle line:84% possibly, besides just actually hearing wh-questions. 00:36:38.450 --> 00:36:42.440 align:middle line:84% So just to make it more specific what I mean, 00:36:42.440 --> 00:36:44.450 align:middle line:84% so here's something which is not true. 00:36:44.450 --> 00:36:47.780 align:middle line:84% If it turned out that every language in which the subject 00:36:47.780 --> 00:36:51.170 align:middle line:84% comes before the verb is a language with wh-movement-- 00:36:51.170 --> 00:36:52.260 align:middle line:90% now, that's false. 00:36:52.260 --> 00:36:55.400 align:middle line:84% So Mandarin is a language where subject comes before the verb, 00:36:55.400 --> 00:36:56.690 align:middle line:90% and it has wh-in-situ. 00:36:56.690 --> 00:36:59.053 align:middle line:84% But imagine that it turned out that every language where 00:36:59.053 --> 00:37:01.220 align:middle line:84% the subject comes before the verb is a language with 00:37:01.220 --> 00:37:03.860 align:middle line:84% wh-movement. , Then an English-speaking baby 00:37:03.860 --> 00:37:05.892 align:middle line:84% wouldn't have to hear any wh-questions at all. 00:37:05.892 --> 00:37:07.850 align:middle line:84% They'd just have to hear that the subject comes 00:37:07.850 --> 00:37:10.017 align:middle line:84% before the verb, and then they would be like, oh, it 00:37:10.017 --> 00:37:11.630 align:middle line:90% must be a wh-movement language. 00:37:11.630 --> 00:37:13.970 align:middle line:84% So if it were possible to connect the fact that English 00:37:13.970 --> 00:37:16.580 align:middle line:84% has wh-movement to some other fact about English, 00:37:16.580 --> 00:37:19.400 align:middle line:84% then babies would have other sources of data. 00:37:19.400 --> 00:37:21.530 align:middle line:84% Now, like I say, that theory that I just 00:37:21.530 --> 00:37:23.820 align:middle line:84% offered as a toy theory, that's not the right theory. 00:37:23.820 --> 00:37:25.945 align:middle line:84% So we have to work harder to find the right theory. 00:37:25.945 --> 00:37:29.040 align:middle line:84% But that's the kind of theory that would give you more data, 00:37:29.040 --> 00:37:31.670 align:middle line:90% yeah Yeah? 00:37:31.670 --> 00:37:35.300 align:middle line:84% AUDIENCE: So you said babies never consider that it could be 00:37:35.300 --> 00:37:39.083 align:middle line:84% [INAUDIBLE] oh, a maximum of two-- 00:37:39.083 --> 00:37:40.000 align:middle line:90% NORVIN RICHARDS: Yeah. 00:37:40.000 --> 00:37:41.560 align:middle line:84% AUDIENCE: --that move to the left. 00:37:41.560 --> 00:37:44.240 align:middle line:84% But how do we know that it's not just like-- 00:37:44.240 --> 00:37:47.740 align:middle line:84% that's not the data that the baby would get? 00:37:47.740 --> 00:37:53.460 align:middle line:84% A baby that has spent eight or nine months just taking 00:37:53.460 --> 00:37:54.433 align:middle line:90% in sound. 00:37:54.433 --> 00:37:55.350 align:middle line:90% NORVIN RICHARDS: Data? 00:37:55.350 --> 00:37:57.300 align:middle line:84% AUDIENCE: And before they even speak a sound. 00:37:57.300 --> 00:37:58.217 align:middle line:90% NORVIN RICHARDS: Yeah. 00:37:58.217 --> 00:38:01.570 align:middle line:84% AUDIENCE: So how do we know that it's just like, 00:38:01.570 --> 00:38:05.050 align:middle line:84% oh, this was not part of what was 00:38:05.050 --> 00:38:07.730 align:middle line:90% presented as data to the baby? 00:38:07.730 --> 00:38:09.520 align:middle line:84% NORVIN RICHARDS: So I don't know if this 00:38:09.520 --> 00:38:11.228 align:middle line:84% is going to address your question or not, 00:38:11.228 --> 00:38:13.300 align:middle line:90% but let me try. 00:38:13.300 --> 00:38:17.680 align:middle line:84% No matter how long we let the baby spend learning Bulgarian, 00:38:17.680 --> 00:38:20.860 align:middle line:84% no matter how much Bulgarian they hear, 00:38:20.860 --> 00:38:24.160 align:middle line:84% there's some maximum number of words 00:38:24.160 --> 00:38:26.020 align:middle line:90% they heard in a wh-question. 00:38:26.020 --> 00:38:30.040 align:middle line:84% It might not be two-- it might be three, it might be four-- 00:38:30.040 --> 00:38:32.230 align:middle line:90% but it's some number. 00:38:32.230 --> 00:38:36.940 align:middle line:84% And so you might have expected that there would be variation 00:38:36.940 --> 00:38:41.290 align:middle line:84% among Bulgarians because any number, whatever the number is, 00:38:41.290 --> 00:38:43.780 align:middle line:84% it doesn't uniquely pick out a grammar. 00:38:43.780 --> 00:38:46.810 align:middle line:84% There's always, actually, infinitely many possible 00:38:46.810 --> 00:38:47.380 align:middle line:90% grammars. 00:38:47.380 --> 00:38:49.660 align:middle line:84% So if the number that they heard was n, 00:38:49.660 --> 00:38:51.940 align:middle line:84% if that's the maximum number, then it 00:38:51.940 --> 00:38:53.140 align:middle line:90% could be "Move them all." 00:38:53.140 --> 00:38:55.255 align:middle line:84% It could be "Move them all, up to a maximum of n." 00:38:55.255 --> 00:38:57.940 align:middle line:84% It could be "Move them all up to a maximum of n plus 1," 00:38:57.940 --> 00:38:59.560 align:middle line:90% or n plus 2, or plus 3. 00:38:59.560 --> 00:39:01.990 align:middle line:84% Yeah, it could be, again, infinitely 00:39:01.990 --> 00:39:03.100 align:middle line:90% many possible grammars. 00:39:03.100 --> 00:39:05.253 align:middle line:84% And so you might have hoped, you might 00:39:05.253 --> 00:39:06.670 align:middle line:84% have expected, that there would be 00:39:06.670 --> 00:39:09.550 align:middle line:84% lots of different adult Bulgarian grammars. 00:39:09.550 --> 00:39:11.680 align:middle line:84% But there aren't; there's only one. 00:39:11.680 --> 00:39:14.800 align:middle line:90% It's "move them all." 00:39:14.800 --> 00:39:17.132 align:middle line:84% Is that getting at your question? 00:39:17.132 --> 00:39:18.190 align:middle line:90% AUDIENCE: Probably, yeah. 00:39:18.190 --> 00:39:19.940 align:middle line:84% NORVIN RICHARDS: So it just doesn't matter 00:39:19.940 --> 00:39:21.640 align:middle line:90% how much data the babies have. 00:39:21.640 --> 00:39:24.490 align:middle line:84% It's never enough to uniquely determine 00:39:24.490 --> 00:39:26.500 align:middle line:90% what the grammar ought to be. 00:39:26.500 --> 00:39:29.980 align:middle line:84% It's got to be the data plus something in the baby's head 00:39:29.980 --> 00:39:34.570 align:middle line:84% that says, the only options are don't move any of them, 00:39:34.570 --> 00:39:36.700 align:middle line:84% move one of them, or move them all. 00:39:36.700 --> 00:39:40.900 align:middle line:84% Those three are the only options because no matter how much data 00:39:40.900 --> 00:39:43.240 align:middle line:84% they have, they'll never uniquely determine 00:39:43.240 --> 00:39:46.700 align:middle line:84% the right data, the right grammar. 00:39:46.700 --> 00:39:47.540 align:middle line:90% Yeah? 00:39:47.540 --> 00:39:49.580 align:middle line:84% AUDIENCE: I don't know a lot about this topic, 00:39:49.580 --> 00:39:52.202 align:middle line:84% but isn't there some merit in studying 00:39:52.202 --> 00:39:55.646 align:middle line:84% constructed languages and the wh-movements that exist there? 00:39:55.646 --> 00:39:58.030 align:middle line:84% Because then you have a lot of liberty 00:39:58.030 --> 00:40:00.743 align:middle line:84% and get to put it wherever you want, I would imagine? 00:40:00.743 --> 00:40:01.660 align:middle line:90% NORVIN RICHARDS: Yeah. 00:40:01.660 --> 00:40:05.246 align:middle line:84% AUDIENCE: And so you could ask, is that consistent 00:40:05.246 --> 00:40:08.485 align:middle line:84% with the order in any [INAUDIBLE].. 00:40:08.485 --> 00:40:10.610 align:middle line:84% NORVIN RICHARDS: So I don't know if this is getting 00:40:10.610 --> 00:40:13.310 align:middle line:84% at your question either, and you should stop me if it doesn't. 00:40:13.310 --> 00:40:14.720 align:middle line:84% No, you should just let me ramble on, 00:40:14.720 --> 00:40:16.303 align:middle line:84% and then you should tell me afterwards 00:40:16.303 --> 00:40:17.870 align:middle line:90% whether it did or not. 00:40:17.870 --> 00:40:19.410 align:middle line:90% That's what you should do. 00:40:19.410 --> 00:40:22.580 align:middle line:84% There are experiments in which people 00:40:22.580 --> 00:40:25.700 align:middle line:84% try to find out-- in which people 00:40:25.700 --> 00:40:27.170 align:middle line:90% ask people to learn languages. 00:40:27.170 --> 00:40:28.940 align:middle line:84% So they'll give people languages, 00:40:28.940 --> 00:40:32.030 align:middle line:84% and they'll find out things like, if I give you 00:40:32.030 --> 00:40:34.940 align:middle line:84% a language that is a possible human language, 00:40:34.940 --> 00:40:37.580 align:middle line:84% are you faster at learning that kind of language 00:40:37.580 --> 00:40:39.680 align:middle line:84% than you would be if I give you a language that's 00:40:39.680 --> 00:40:41.298 align:middle line:90% not a possible human language? 00:40:41.298 --> 00:40:43.340 align:middle line:84% So if I give you a language in which the rule, is 00:40:43.340 --> 00:40:46.250 align:middle line:84% move all your wh-phrases, how are you at learning that 00:40:46.250 --> 00:40:47.690 align:middle line:84% as opposed to a language in which 00:40:47.690 --> 00:40:49.610 align:middle line:84% the rule is, move all your wh-phrases up 00:40:49.610 --> 00:40:51.110 align:middle line:90% to a maximum of two? 00:40:51.110 --> 00:40:54.020 align:middle line:84% And I don't know whether that kind of experiment 00:40:54.020 --> 00:40:55.770 align:middle line:84% has been done for this particular problem. 00:40:55.770 --> 00:40:57.520 align:middle line:84% But what we would hope is that, yeah, it's 00:40:57.520 --> 00:40:59.000 align:middle line:84% harder to learn the languages where 00:40:59.000 --> 00:41:01.370 align:middle line:84% it's up to a maximum of two, the languages-- 00:41:01.370 --> 00:41:03.500 align:middle line:84% people do do this kind of experiment. 00:41:03.500 --> 00:41:05.118 align:middle line:84% The other kind of natural experiment 00:41:05.118 --> 00:41:07.160 align:middle line:84% that's been done-- this is a lot more anecdotal-- 00:41:07.160 --> 00:41:11.000 align:middle line:84% there's, what are people like this called now? 00:41:11.000 --> 00:41:17.660 align:middle line:84% There's a savant, so someone whose mental processes 00:41:17.660 --> 00:41:21.140 align:middle line:90% are mostly very slow. 00:41:21.140 --> 00:41:25.310 align:middle line:84% He needs a lot help with daily life. 00:41:25.310 --> 00:41:27.560 align:middle line:84% I think he does live by himself, but he has caregivers 00:41:27.560 --> 00:41:29.490 align:middle line:90% who come and help him out. 00:41:29.490 --> 00:41:31.540 align:middle line:84% But he's a genius at learning languages. 00:41:31.540 --> 00:41:33.773 align:middle line:84% And so I don't know his real name, 00:41:33.773 --> 00:41:35.690 align:middle line:84% but his name in the literature is Christopher. 00:41:35.690 --> 00:41:37.550 align:middle line:90% He's British. 00:41:37.550 --> 00:41:41.480 align:middle line:84% And Christopher loves learning languages, does it for fun. 00:41:41.480 --> 00:41:43.730 align:middle line:84% You can give him grammars, and he will learn languages 00:41:43.730 --> 00:41:44.300 align:middle line:90% immediately. 00:41:44.300 --> 00:41:46.160 align:middle line:84% It takes him a day, and suddenly, he 00:41:46.160 --> 00:41:47.350 align:middle line:90% can read and write. 00:41:47.350 --> 00:41:51.470 align:middle line:84% And he's not good at speaking the languages 00:41:51.470 --> 00:41:54.050 align:middle line:84% partly because he learns by reading books 00:41:54.050 --> 00:41:56.150 align:middle line:84% but he can learn to read and write languages 00:41:56.150 --> 00:41:57.710 align:middle line:90% extremely quickly. 00:41:57.710 --> 00:41:59.240 align:middle line:84% And they have done some experiments 00:41:59.240 --> 00:42:01.010 align:middle line:84% like this with Christopher where you give Christopher 00:42:01.010 --> 00:42:03.230 align:middle line:84% languages that are not possible languages to learn, 00:42:03.230 --> 00:42:06.320 align:middle line:84% and he can't do that, it turns out. 00:42:06.320 --> 00:42:09.230 align:middle line:84% So he can learn languages as long 00:42:09.230 --> 00:42:12.450 align:middle line:84% as they are languages that the rules of universal grammar 00:42:12.450 --> 00:42:12.950 align:middle line:90% allow. 00:42:12.950 --> 00:42:14.408 align:middle line:84% Now, again, I don't know if they've 00:42:14.408 --> 00:42:17.942 align:middle line:84% tried "Move all your wh-phrases up to a maximum of two" 00:42:17.942 --> 00:42:18.650 align:middle line:90% with Christopher. 00:42:18.650 --> 00:42:21.200 align:middle line:84% But that's another kind of experiment that's been done. 00:42:21.200 --> 00:42:21.770 align:middle line:90% Yeah? 00:42:21.770 --> 00:42:23.660 align:middle line:84% AUDIENCE: How many languages will Christopher 00:42:23.660 --> 00:42:26.322 align:middle line:84% have learned by the time [INAUDIBLE]?? 00:42:26.322 --> 00:42:28.990 align:middle line:84% Because there's maybe a possibility 00:42:28.990 --> 00:42:31.690 align:middle line:84% that Christopher unconsciously figured it out 00:42:31.690 --> 00:42:35.520 align:middle line:84% that it has to be none, one, or all. 00:42:35.520 --> 00:42:37.980 align:middle line:84% That's not from Christopher's experience-- 00:42:37.980 --> 00:42:39.480 align:middle line:84% NORVIN RICHARDS: His wide experience 00:42:39.480 --> 00:42:41.430 align:middle line:90% with previous languages? 00:42:41.430 --> 00:42:45.990 align:middle line:84% So all I can say is that they have given him languages-- 00:42:45.990 --> 00:42:47.530 align:middle line:90% they have worried about this. 00:42:47.530 --> 00:42:49.800 align:middle line:84% So they have given him languages-- thank you 00:42:49.800 --> 00:42:51.410 align:middle line:90% for figuring out the doors. 00:42:51.410 --> 00:42:54.150 align:middle line:84% [LAUGHS] They have given him languages 00:42:54.150 --> 00:42:58.710 align:middle line:84% that differ from the languages he's learned in some regard. 00:42:58.710 --> 00:43:00.300 align:middle line:84% Most of the languages he's learned 00:43:00.300 --> 00:43:02.170 align:middle line:90% are various languages of Europe. 00:43:02.170 --> 00:43:05.440 align:middle line:90% He's done like 20 or 30 by now. 00:43:05.440 --> 00:43:09.810 align:middle line:84% And so they have given him languages that differ from 00:43:09.810 --> 00:43:12.570 align:middle line:84% the languages that he's known to find out whether he can learn 00:43:12.570 --> 00:43:14.910 align:middle line:84% them just as quickly as languages that-- 00:43:14.910 --> 00:43:17.160 align:middle line:84% because they also get, when they do these experiments, 00:43:17.160 --> 00:43:19.830 align:middle line:84% they'll give him a language that does obey some rule 00:43:19.830 --> 00:43:22.635 align:middle line:84% of universal grammar and another language that doesn't. 00:43:22.635 --> 00:43:24.510 align:middle line:84% But you're absolutely right, that's something 00:43:24.510 --> 00:43:25.635 align:middle line:90% they have to watch out for. 00:43:25.635 --> 00:43:28.290 align:middle line:84% And all I know is that they have been careful about that. 00:43:28.290 --> 00:43:30.770 align:middle line:84% Whether they've been careful enough, I can't swear. 00:43:30.770 --> 00:43:33.232 align:middle line:90% Yeah, that's a good point. 00:43:33.232 --> 00:43:34.940 align:middle line:84% Christopher loves doing this, by the way. 00:43:34.940 --> 00:43:37.232 align:middle line:84% This is not a hardship they're imposing on Christopher. 00:43:37.232 --> 00:43:39.890 align:middle line:84% This is one of his big joys in life, is learning languages. 00:43:39.890 --> 00:43:45.285 align:middle line:90% Yeah, OK. 00:43:45.285 --> 00:43:46.160 align:middle line:90% [? Let me ?] back in? 00:43:46.160 --> 00:43:47.360 align:middle line:90% Yes. 00:43:47.360 --> 00:43:49.650 align:middle line:90% Yeah, questions about this? 00:43:49.650 --> 00:43:53.840 align:middle line:90% OK, all right. 00:43:53.840 --> 00:43:57.350 align:middle line:84% All right, so that's it for wh-questions, at least for now. 00:43:57.350 --> 00:44:00.080 align:middle line:84% Are there any wh-question questions 00:44:00.080 --> 00:44:03.547 align:middle line:84% before we put that aside and go on to something else? 00:44:03.547 --> 00:44:04.617 align:middle line:90% OK. 00:44:04.617 --> 00:44:06.950 align:middle line:84% We have spent a lot of time talking about the projection 00:44:06.950 --> 00:44:07.910 align:middle line:90% principle-- 00:44:07.910 --> 00:44:11.960 align:middle line:84% this idea that if you have a head, and it selects something, 00:44:11.960 --> 00:44:14.960 align:middle line:84% that thing needs to be its sister. 00:44:14.960 --> 00:44:17.570 align:middle line:84% Then, I've incautiously put the verb "put" 00:44:17.570 --> 00:44:20.245 align:middle line:84% in this example where, I don't know, it needs two sisters, 00:44:20.245 --> 00:44:21.620 align:middle line:84% or we need to say something else. 00:44:21.620 --> 00:44:24.140 align:middle line:84% Anyway, there are strict locality restrictions 00:44:24.140 --> 00:44:26.780 align:middle line:84% on the relationship between the head and the thing 00:44:26.780 --> 00:44:28.220 align:middle line:90% that it selects. 00:44:28.220 --> 00:44:31.070 align:middle line:84% So we have lexical entries for verbs like "devour," 00:44:31.070 --> 00:44:34.340 align:middle line:84% and, "put," and "faint," that say things like "'devour' needs 00:44:34.340 --> 00:44:37.095 align:middle line:84% a sister that's a noun phrase"-- and "'put' needs to combine 00:44:37.095 --> 00:44:39.470 align:middle line:84% with a noun phrase and also with a prepositional phrase," 00:44:39.470 --> 00:44:41.262 align:middle line:90% and "'faint' is intransitive." 00:44:41.262 --> 00:44:43.220 align:middle line:84% It had better not combine with anything at all. 00:44:43.220 --> 00:44:47.600 align:middle line:84% Yeah, those are all things that we've been saying. 00:44:47.600 --> 00:44:49.790 align:middle line:84% And what we've been saying is, if you 00:44:49.790 --> 00:44:51.860 align:middle line:84% have a head that doesn't select for a sister, 00:44:51.860 --> 00:44:53.790 align:middle line:84% well, then, it doesn't have a sister-- 00:44:53.790 --> 00:44:54.800 align:middle line:90% the end. 00:44:54.800 --> 00:44:58.340 align:middle line:84% So "he fainted," "faint," it's an intransitive verb. 00:44:58.340 --> 00:45:00.260 align:middle line:84% Doesn't need a sister, doesn't get one, 00:45:00.260 --> 00:45:02.833 align:middle line:90% and that's the end, OK? 00:45:02.833 --> 00:45:05.250 align:middle line:84% Unless you add an adjunct, but it doesn't get an argument. 00:45:05.250 --> 00:45:07.870 align:middle line:90% 00:45:07.870 --> 00:45:10.900 align:middle line:84% But there's nothing comparable in subject position. 00:45:10.900 --> 00:45:12.390 align:middle line:84% So when we talk about verbs being 00:45:12.390 --> 00:45:16.380 align:middle line:84% transitive or intransitive, we're talking about objects. 00:45:16.380 --> 00:45:17.790 align:middle line:84% So people have been talking about 00:45:17.790 --> 00:45:19.230 align:middle line:84% transitive and intransitive verbs, 00:45:19.230 --> 00:45:20.903 align:middle line:90% literally, for centuries. 00:45:20.903 --> 00:45:22.320 align:middle line:84% It's an old observation that there 00:45:22.320 --> 00:45:24.903 align:middle line:84% are verbs that need objects and verbs that don't need objects. 00:45:24.903 --> 00:45:27.060 align:middle line:84% But there's nothing comparable in subject position. 00:45:27.060 --> 00:45:28.643 align:middle line:84% That is, people don't talk about verbs 00:45:28.643 --> 00:45:31.020 align:middle line:84% that need subjects and verbs that don't need subjects. 00:45:31.020 --> 00:45:33.990 align:middle line:90% Verbs just always have subjects. 00:45:33.990 --> 00:45:40.110 align:middle line:84% So and sometimes, the subjects are a little sketchy. 00:45:40.110 --> 00:45:45.000 align:middle line:84% I mean, so we say things like "It rained," and it's not 00:45:45.000 --> 00:45:47.520 align:middle line:84% too clear what "it" is supposed to be. 00:45:47.520 --> 00:45:49.140 align:middle line:90% You can't be anything else. 00:45:49.140 --> 00:45:52.290 align:middle line:84% You don't get to say, "the sky rained," 00:45:52.290 --> 00:45:55.440 align:middle line:90% or "God rained," or whatever. 00:45:55.440 --> 00:45:56.970 align:middle line:90% It sort of needs to be "it." 00:45:56.970 --> 00:45:59.962 align:middle line:90% 00:45:59.962 --> 00:46:02.170 align:middle line:84% And there are other kinds of examples like that, too. 00:46:02.170 --> 00:46:04.770 align:middle line:84% So "It seems that John has died," 00:46:04.770 --> 00:46:08.010 align:middle line:84% if I say that to you, you're not going to say, 00:46:08.010 --> 00:46:10.380 align:middle line:90% "What seems that John has died? 00:46:10.380 --> 00:46:11.760 align:middle line:90% What is 'it'? 00:46:11.760 --> 00:46:14.050 align:middle line:90% What do you mean by 'it' there?" 00:46:14.050 --> 00:46:15.480 align:middle line:90% I don't mean anything at all. 00:46:15.480 --> 00:46:16.938 align:middle line:90% "It" is just there-- 00:46:16.938 --> 00:46:19.230 align:middle line:84% it's as though it's there so that the sentence can have 00:46:19.230 --> 00:46:23.220 align:middle line:84% a subject, even though the subject doesn't 00:46:23.220 --> 00:46:27.340 align:middle line:84% contribute anything to the meaning of the sentence. 00:46:27.340 --> 00:46:33.410 align:middle line:84% So yeah, if you compare-- here's a short, kind of sad story. 00:46:33.410 --> 00:46:36.460 align:middle line:84% So "It squeezed John, and it seems that John has died." 00:46:36.460 --> 00:46:39.420 align:middle line:90% Is anybody here named John? 00:46:39.420 --> 00:46:39.920 align:middle line:90% Good! 00:46:39.920 --> 00:46:41.330 align:middle line:90% [LAUGHTER] 00:46:41.330 --> 00:46:45.260 align:middle line:84% So I should probably have warned you about this early on. 00:46:45.260 --> 00:46:48.650 align:middle line:84% Example sentences in linguistics, I mean, they vary. 00:46:48.650 --> 00:46:51.800 align:middle line:84% Often, you need a name, and John is one of the classics. 00:46:51.800 --> 00:46:54.770 align:middle line:90% Mary is another one. 00:46:54.770 --> 00:46:57.950 align:middle line:84% So the first time I ever taught intro to linguistics, there 00:46:57.950 --> 00:46:59.630 align:middle line:84% was a John and a Mary in the class. 00:46:59.630 --> 00:47:02.780 align:middle line:90% It was very confining. 00:47:02.780 --> 00:47:04.850 align:middle line:84% So since there's no John, I can just 00:47:04.850 --> 00:47:06.590 align:middle line:90% let these example sentences rip. 00:47:06.590 --> 00:47:07.460 align:middle line:90% "It squeezed John. 00:47:07.460 --> 00:47:09.680 align:middle line:90% It seems that John has died." 00:47:09.680 --> 00:47:12.110 align:middle line:84% These two "it"s seem to have a different status. 00:47:12.110 --> 00:47:15.470 align:middle line:84% So with the first one, if I say, "It squeezed John," 00:47:15.470 --> 00:47:17.330 align:middle line:84% if I say that out of the blue, you're 00:47:17.330 --> 00:47:19.520 align:middle line:84% entitled to wonder, "What squeezed John?" 00:47:19.520 --> 00:47:21.320 align:middle line:90% That's a reasonable question. 00:47:21.320 --> 00:47:24.950 align:middle line:84% So it's some machine, or a snake, or something. 00:47:24.950 --> 00:47:26.300 align:middle line:90% Something squeezed John. 00:47:26.300 --> 00:47:29.030 align:middle line:84% But if I say, "it seems that John has died, " 00:47:29.030 --> 00:47:32.510 align:middle line:84% even if you don't care about John at all, 00:47:32.510 --> 00:47:34.100 align:middle line:84% your question is not going to be, 00:47:34.100 --> 00:47:36.140 align:middle line:90% "What seems that John has died?" 00:47:36.140 --> 00:47:37.700 align:middle line:84% That question doesn't make any sense. 00:47:37.700 --> 00:47:40.540 align:middle line:90% 00:47:40.540 --> 00:47:42.665 align:middle line:84% So these two "it"s seem to have a different status. 00:47:42.665 --> 00:47:45.716 align:middle line:90% 00:47:45.716 --> 00:47:49.260 align:middle line:84% There's a classic way of handling this kind of fact, 00:47:49.260 --> 00:47:52.370 align:middle line:84% which is to offer what's called the Extended Projection 00:47:52.370 --> 00:47:55.438 align:middle line:90% Principle, which-- 00:47:55.438 --> 00:47:57.230 align:middle line:84% yeah, so the Extended Projection Principle. 00:47:57.230 --> 00:47:58.772 align:middle line:84% So you know the Projection Principle, 00:47:58.772 --> 00:48:00.860 align:middle line:84% which says that heads, when they select, 00:48:00.860 --> 00:48:03.020 align:middle line:84% they need to select things that are their sisters, 00:48:03.020 --> 00:48:06.360 align:middle line:84% or they need their things they select to be close to them. 00:48:06.360 --> 00:48:08.990 align:middle line:84% The Extended Projection Principle says, there 00:48:08.990 --> 00:48:10.730 align:middle line:90% needs to be something here. 00:48:10.730 --> 00:48:13.700 align:middle line:90% So TP needs to have a specifier. 00:48:13.700 --> 00:48:17.210 align:middle line:84% You can't have a sentence that doesn't have a subject in it. 00:48:17.210 --> 00:48:20.630 align:middle line:84% Calling it the Extended Projection Principle, 00:48:20.630 --> 00:48:23.040 align:middle line:90% this was Noam Chomsky's idea. 00:48:23.040 --> 00:48:28.015 align:middle line:84% And so I feel secure in telling you that it's an obvious hack. 00:48:28.015 --> 00:48:29.390 align:middle line:84% The Extended Projection Principle 00:48:29.390 --> 00:48:32.000 align:middle line:84% has nothing to do with the Projection Principle. 00:48:32.000 --> 00:48:33.740 align:middle line:84% He really just called it that, I think, 00:48:33.740 --> 00:48:35.750 align:middle line:84% in order to paint a big red target on it. 00:48:35.750 --> 00:48:38.720 align:middle line:84% He's like, why the heck does it work this way, 00:48:38.720 --> 00:48:41.630 align:middle line:84% that there has to be something in the specifier of TP? 00:48:41.630 --> 00:48:43.370 align:middle line:84% Linguists, go and try to figure out what 00:48:43.370 --> 00:48:45.290 align:middle line:90% the heck is going on there. 00:48:45.290 --> 00:48:46.790 align:middle line:84% So the Extended Projection Principle 00:48:46.790 --> 00:48:51.077 align:middle line:84% says the specifier of TP-- this is a word I've used before, 00:48:51.077 --> 00:48:53.660 align:middle line:84% and I hope I've defined it, but if I haven't, let me define it 00:48:53.660 --> 00:48:55.490 align:middle line:84% for you-- the specifier of TP, there 00:48:55.490 --> 00:48:57.320 align:middle line:84% has to be something in the specifier of TP. 00:48:57.320 --> 00:48:58.850 align:middle line:90% What's the specifier of TP? 00:48:58.850 --> 00:49:02.360 align:middle line:84% It's a daughter of TP-- so a daughter of the maximum 00:49:02.360 --> 00:49:03.830 align:middle line:90% projection-- 00:49:03.830 --> 00:49:05.930 align:middle line:90% that doesn't have the label T-- 00:49:05.930 --> 00:49:08.630 align:middle line:84% so it's that thing that the red arrow is pointing to-- 00:49:08.630 --> 00:49:11.150 align:middle line:84% and it's not selected, or at least it 00:49:11.150 --> 00:49:13.020 align:middle line:84% doesn't have to be selected, by anything. 00:49:13.020 --> 00:49:15.620 align:middle line:84% So here's another place where there has to be something. 00:49:15.620 --> 00:49:18.710 align:middle line:84% If you have a transitive verb, the transitive verb 00:49:18.710 --> 00:49:21.385 align:middle line:90% has to have a sister, an object. 00:49:21.385 --> 00:49:23.510 align:middle line:84% If you have a verb that selects for something else, 00:49:23.510 --> 00:49:25.580 align:middle line:84% like "depends" selects for a prepositional phrase 00:49:25.580 --> 00:49:28.560 align:middle line:84% headed by "on", and it has to have that prepositional phrase. 00:49:28.560 --> 00:49:29.560 align:middle line:90% Here's another position. 00:49:29.560 --> 00:49:31.430 align:middle line:84% It doesn't seem to have anything to do with selection. 00:49:31.430 --> 00:49:32.888 align:middle line:84% It doesn't matter what the verb is. 00:49:32.888 --> 00:49:35.510 align:middle line:84% There just has to be something in subject position. 00:49:35.510 --> 00:49:38.450 align:middle line:84% So it's called the Extended Projection Principle. 00:49:38.450 --> 00:49:43.400 align:middle line:84% Calling it a principle is dignifying it too much. 00:49:43.400 --> 00:49:47.120 align:middle line:84% And what we're seeing is that in sentences where there isn't 00:49:47.120 --> 00:49:50.120 align:middle line:84% anything else to satisfy the Extended Projection Principle 00:49:50.120 --> 00:49:52.370 align:middle line:90% with, then you insert "it." 00:49:52.370 --> 00:49:54.950 align:middle line:90% You can insert "it." 00:49:54.950 --> 00:49:57.930 align:middle line:84% That's what you're doing in "It seems that John has died." 00:49:57.930 --> 00:50:00.233 align:middle line:84% That is, you have this subject position. 00:50:00.233 --> 00:50:01.650 align:middle line:84% There needs to be something there, 00:50:01.650 --> 00:50:04.580 align:middle line:84% and so you put this meaningless "it" there. 00:50:04.580 --> 00:50:06.710 align:middle line:84% This is why you can't ask questions like, "What 00:50:06.710 --> 00:50:08.090 align:middle line:90% seems that John has died?" 00:50:08.090 --> 00:50:10.700 align:middle line:84% Because there's nothing meaningful there. 00:50:10.700 --> 00:50:15.573 align:middle line:84% That meaningless version of "it" is called an expletive. 00:50:15.573 --> 00:50:17.240 align:middle line:84% There are other names for it, but that's 00:50:17.240 --> 00:50:19.680 align:middle line:90% what we usually call it. 00:50:19.680 --> 00:50:22.050 align:middle line:90% So it doesn't mean anything. 00:50:22.050 --> 00:50:24.950 align:middle line:84% It's just there to satisfy the EPP, the Extended Projection 00:50:24.950 --> 00:50:28.100 align:middle line:90% Principle. 00:50:28.100 --> 00:50:31.470 align:middle line:84% I think I may say this on a later slide. 00:50:31.470 --> 00:50:35.760 align:middle line:84% Yeah, we've just gone through saying 00:50:35.760 --> 00:50:38.550 align:middle line:90% languages vary in various ways. 00:50:38.550 --> 00:50:40.290 align:middle line:84% This is a place where languages vary. 00:50:40.290 --> 00:50:42.000 align:middle line:84% English has the Extended Projection Principle. 00:50:42.000 --> 00:50:43.050 align:middle line:90% There are other languages. 00:50:43.050 --> 00:50:44.640 align:middle line:84% French has the Extended Projection principle 00:50:44.640 --> 00:50:45.610 align:middle line:90% There are some others. 00:50:45.610 --> 00:50:48.193 align:middle line:84% But there are many languages-- in fact, maybe most languages-- 00:50:48.193 --> 00:50:49.630 align:middle line:90% that don't have this. 00:50:49.630 --> 00:50:52.230 align:middle line:84% So there's languages that are perfectly happy to start 00:50:52.230 --> 00:50:53.770 align:middle line:90% the sentence with the verb. 00:50:53.770 --> 00:50:55.958 align:middle line:84% So if you're working on a language, 00:50:55.958 --> 00:50:58.500 align:middle line:84% yeah, if you speak Spanish, then you're thinking to yourself, 00:50:58.500 --> 00:51:00.060 align:middle line:84% wait, it's fine to say things like, 00:51:00.060 --> 00:51:03.690 align:middle line:84% "Seems that John has died" in lots of languages of the world. 00:51:03.690 --> 00:51:06.770 align:middle line:90% 00:51:06.770 --> 00:51:09.270 align:middle line:84% So yeah, here's another point of cross-linguistic variation. 00:51:09.270 --> 00:51:11.100 align:middle line:84% There are languages like English that have 00:51:11.100 --> 00:51:12.802 align:middle line:90% this interesting restriction. 00:51:12.802 --> 00:51:16.580 align:middle line:90% 00:51:16.580 --> 00:51:19.028 align:middle line:90% OK, now here's another-- 00:51:19.028 --> 00:51:21.320 align:middle line:84% I'm telling you about the Extended Protection Principle 00:51:21.320 --> 00:51:24.660 align:middle line:84% because I want to introduce you to another kind of movement. 00:51:24.660 --> 00:51:26.720 align:middle line:84% So we've talked about movement, and I 00:51:26.720 --> 00:51:29.330 align:middle line:90% want to show you another kind. 00:51:29.330 --> 00:51:31.160 align:middle line:90% Here's another kind. 00:51:31.160 --> 00:51:34.070 align:middle line:84% We can say things like "The snake squeezed John," 00:51:34.070 --> 00:51:36.200 align:middle line:84% or "I put the kumquats in a bowl." 00:51:36.200 --> 00:51:40.860 align:middle line:84% And then, so let's concentrate on "The snake squeezed John." 00:51:40.860 --> 00:51:43.800 align:middle line:84% "Squeeze" looks like it's a transitive verb, 00:51:43.800 --> 00:51:46.080 align:middle line:84% or at least it has a life as a transitive verb. 00:51:46.080 --> 00:51:48.570 align:middle line:84% "John" is being selected by "squeeze." 00:51:48.570 --> 00:51:51.990 align:middle line:84% It's the thing that's getting squeezed. 00:51:51.990 --> 00:51:54.720 align:middle line:84% Or similarly, in "I put the kumquats in the bowl," 00:51:54.720 --> 00:51:57.700 align:middle line:84% "put" is a verb that selects for an object. 00:51:57.700 --> 00:51:59.680 align:middle line:84% It also selects for a prepositional phrase. 00:51:59.680 --> 00:52:01.540 align:middle line:90% That's what we've been saying. 00:52:01.540 --> 00:52:04.420 align:middle line:84% But then, there are these other forms, these passives-- 00:52:04.420 --> 00:52:08.200 align:middle line:84% "John was squeezed," and "The kumquats were put in a bowl." 00:52:08.200 --> 00:52:11.770 align:middle line:84% And here, we're in a similar situation 00:52:11.770 --> 00:52:14.440 align:middle line:84% to the situation we were in when I first showed you 00:52:14.440 --> 00:52:16.060 align:middle line:90% "What did Mary write?" 00:52:16.060 --> 00:52:25.100 align:middle line:84% where I was trying to say "What did Mary put in the bowl?" 00:52:25.100 --> 00:52:28.822 align:middle line:84% So with "What did Mary put in the bowl?" 00:52:28.822 --> 00:52:30.280 align:middle line:84% I was saying, yeah, "put," we think 00:52:30.280 --> 00:52:32.232 align:middle line:84% "put" has to combine with a noun phrase. 00:52:32.232 --> 00:52:33.190 align:middle line:90% It's a transitive verb. 00:52:33.190 --> 00:52:35.080 align:middle line:84% You can't say, "Mary put in the bowl." 00:52:35.080 --> 00:52:37.220 align:middle line:90% That's not a sentence. 00:52:37.220 --> 00:52:40.610 align:middle line:84% Why is it OK to have "put" with no noun phrase after it here? 00:52:40.610 --> 00:52:44.050 align:middle line:84% Well, it's because "what" started out here and moved over 00:52:44.050 --> 00:52:46.000 align:middle line:90% to there. 00:52:46.000 --> 00:52:48.268 align:middle line:84% For "the kumquats were put in a bowl," 00:52:48.268 --> 00:52:49.810 align:middle line:84% we're going to say something similar. 00:52:49.810 --> 00:52:52.240 align:middle line:84% "Put" is selecting for an object. 00:52:52.240 --> 00:52:54.520 align:middle line:84% In this case, it's the kumquats, and the kumquats 00:52:54.520 --> 00:52:57.857 align:middle line:84% are moving out of that position into another position. 00:52:57.857 --> 00:52:59.440 align:middle line:84% What's the position they're moving to? 00:52:59.440 --> 00:53:00.790 align:middle line:90% Well, they're not wh-moving. 00:53:00.790 --> 00:53:02.650 align:middle line:90% They're not wh-words. 00:53:02.650 --> 00:53:04.010 align:middle line:90% This isn't a question. 00:53:04.010 --> 00:53:07.130 align:middle line:84% This is another kind of movement. 00:53:07.130 --> 00:53:09.250 align:middle line:84% So the same reasoning that prompted 00:53:09.250 --> 00:53:11.680 align:middle line:84% this to posit wh-movement in order 00:53:11.680 --> 00:53:14.590 align:middle line:84% to save our beliefs about how selection works 00:53:14.590 --> 00:53:19.430 align:middle line:84% leads us to suspect that there's movement going on here as well. 00:53:19.430 --> 00:53:22.540 align:middle line:84% So we'll start off with "Was squeezed John" 00:53:22.540 --> 00:53:26.470 align:middle line:84% or "Were put kumquats in the bowl." 00:53:26.470 --> 00:53:30.250 align:middle line:84% And then, the EPP is going to say, oh, there's no subject. 00:53:30.250 --> 00:53:33.040 align:middle line:84% So what happens with the passive is that you get rid 00:53:33.040 --> 00:53:34.630 align:middle line:90% of the subject. 00:53:34.630 --> 00:53:37.120 align:middle line:84% "Was squeezed John," you've gotten rid 00:53:37.120 --> 00:53:38.890 align:middle line:84% of the snake that was squeezing John. 00:53:38.890 --> 00:53:42.312 align:middle line:84% Or "Were put in the bowl," "Were put the kumquats in the bowl," 00:53:42.312 --> 00:53:44.020 align:middle line:84% you've gotten rid of Mary who was putting 00:53:44.020 --> 00:53:45.062 align:middle line:90% the kumquats in the bowl. 00:53:45.062 --> 00:53:47.410 align:middle line:84% And then, the EPP says, oh, no, wait, 00:53:47.410 --> 00:53:49.490 align:middle line:84% we need something in subject position. 00:53:49.490 --> 00:53:53.180 align:middle line:84% And so we move-- in this case "John," or in the other case, 00:53:53.180 --> 00:53:55.147 align:middle line:84% "the kumquats"-- into subject positions. 00:53:55.147 --> 00:53:56.230 align:middle line:90% So there can be a subject. 00:53:56.230 --> 00:53:58.900 align:middle line:90% So the EPP can be satisfied. 00:53:58.900 --> 00:54:01.610 align:middle line:84% It's a way of talking about these kinds of examples. 00:54:01.610 --> 00:54:05.170 align:middle line:84% So this is a new kind of movement driven by the EPP. 00:54:05.170 --> 00:54:08.710 align:middle line:84% It's called-- sometimes called NP-movement. 00:54:08.710 --> 00:54:11.440 align:middle line:84% It's been called other things, but I'll call it NP-movement. 00:54:11.440 --> 00:54:13.870 align:middle line:84% It always seems to be movement of NPs. 00:54:13.870 --> 00:54:16.600 align:middle line:84% So you take an NP like "John," and you move it, 00:54:16.600 --> 00:54:19.480 align:middle line:84% in this case, out of the position 00:54:19.480 --> 00:54:22.870 align:middle line:84% as the sister of "squeeze," where it was selected, 00:54:22.870 --> 00:54:25.540 align:middle line:84% into subject position so that the sentence can 00:54:25.540 --> 00:54:28.570 align:middle line:84% have a subject, so that TP can have a specifier. 00:54:28.570 --> 00:54:31.010 align:middle line:90% It seems to be what's happening. 00:54:31.010 --> 00:54:34.780 align:middle line:84% Let me give you some reasons to take that analysis seriously. 00:54:34.780 --> 00:54:36.790 align:middle line:84% But first, is the analysis clear? 00:54:36.790 --> 00:54:41.020 align:middle line:84% That's what we think is happening. 00:54:41.020 --> 00:54:41.740 align:middle line:90% Oh, sorry. 00:54:41.740 --> 00:54:43.910 align:middle line:84% First, let me show you another kind of NP-movement. 00:54:43.910 --> 00:54:46.720 align:middle line:84% So those first ones were passives. 00:54:46.720 --> 00:54:47.630 align:middle line:90% Here's another one. 00:54:47.630 --> 00:54:51.100 align:middle line:84% So in "It seems that John is sick," 00:54:51.100 --> 00:54:53.630 align:middle line:84% there's an "it" at the beginning of the sentence 00:54:53.630 --> 00:54:55.990 align:middle line:84% which is the expletive kind of "it." 00:54:55.990 --> 00:54:58.157 align:middle line:84% So it's the kind of "it" that doesn't mean anything. 00:54:58.157 --> 00:55:00.282 align:middle line:84% I think it was one of the first kinds of expletives 00:55:00.282 --> 00:55:01.210 align:middle line:90% that I showed you. 00:55:01.210 --> 00:55:03.970 align:middle line:84% Can't say things like, "What seems that John is sick?" 00:55:03.970 --> 00:55:06.020 align:middle line:84% So that "q" doesn't mean anything. 00:55:06.020 --> 00:55:10.090 align:middle line:84% It's just there so that TP can have a specifier. 00:55:10.090 --> 00:55:14.525 align:middle line:84% "John" is starting off as the subject of "be sick." 00:55:14.525 --> 00:55:15.900 align:middle line:84% And in the first sentence, "John" 00:55:15.900 --> 00:55:18.030 align:middle line:84% is staying there in subject position for "be sick." 00:55:18.030 --> 00:55:21.510 align:middle line:84% But in the second sentence, "John" appears to be moving. 00:55:21.510 --> 00:55:23.377 align:middle line:84% So it's an analysis, and I'm going 00:55:23.377 --> 00:55:25.710 align:middle line:84% to give you some reasons to take this analysis seriously 00:55:25.710 --> 00:55:26.340 align:middle line:90% in a second. 00:55:26.340 --> 00:55:29.100 align:middle line:84% We're going to say, yeah, "John" is still the subject of "be 00:55:29.100 --> 00:55:30.640 align:middle line:90% sick" in the second example. 00:55:30.640 --> 00:55:35.040 align:middle line:84% But then, he moves to satisfy the EPP in the higher clause. 00:55:35.040 --> 00:55:37.530 align:middle line:84% So the higher verb needs to have a subject, 00:55:37.530 --> 00:55:39.690 align:middle line:84% and here, for some reason, we've elected 00:55:39.690 --> 00:55:41.130 align:middle line:90% not to merge an expletive. 00:55:41.130 --> 00:55:43.920 align:middle line:84% And so instead, we're moving "John" 00:55:43.920 --> 00:55:48.245 align:middle line:84% so that "John" can satisfy the EPP upstairs. 00:55:48.245 --> 00:55:49.620 align:middle line:84% Now, let me give you some reasons 00:55:49.620 --> 00:55:52.140 align:middle line:90% to take this analysis seriously. 00:55:52.140 --> 00:55:55.200 align:middle line:84% So here is an argument for the existence of NP-movement, 00:55:55.200 --> 00:55:57.253 align:middle line:90% and it involves idioms. 00:55:57.253 --> 00:55:58.920 align:middle line:84% So let's talk for a second about idioms. 00:55:58.920 --> 00:56:01.530 align:middle line:84% Idioms are always fun to talk about. 00:56:01.530 --> 00:56:07.080 align:middle line:84% Idioms are combinations of words that 00:56:07.080 --> 00:56:10.410 align:middle line:84% don't have to mean what they appear to literally mean. 00:56:10.410 --> 00:56:14.130 align:middle line:84% So "kick the bucket" and "buy the farm" 00:56:14.130 --> 00:56:17.080 align:middle line:90% are both idioms that mean "die." 00:56:17.080 --> 00:56:19.530 align:middle line:84% So if you say "John bought the farm"-- 00:56:19.530 --> 00:56:22.973 align:middle line:84% continuing to pick on John, since there are no Johns here-- 00:56:22.973 --> 00:56:24.390 align:middle line:84% if you say "John bought the farm," 00:56:24.390 --> 00:56:27.480 align:middle line:84% that could mean John purchased a farm. 00:56:27.480 --> 00:56:29.460 align:middle line:90% It has its literal meaning. 00:56:29.460 --> 00:56:32.160 align:middle line:84% But it can also mean, "John died." 00:56:32.160 --> 00:56:36.120 align:middle line:84% Or, "John kicked the bucket" can literally mean, you know, 00:56:36.120 --> 00:56:37.440 align:middle line:90% John kicked a bucket. 00:56:37.440 --> 00:56:40.230 align:middle line:84% Yeah, but it can also mean that John died. 00:56:40.230 --> 00:56:41.530 align:middle line:90% Do people have these idioms? 00:56:41.530 --> 00:56:43.130 align:middle line:84% Are these things that you can say? 00:56:43.130 --> 00:56:44.880 align:middle line:84% AUDIENCE: I've never heard "buy the farm." 00:56:44.880 --> 00:56:46.020 align:middle line:84% NORVIN RICHARDS: You've never heard "buy the farm"? 00:56:46.020 --> 00:56:47.987 align:middle line:84% Has anybody heard "buy the farm"? 00:56:47.987 --> 00:56:48.570 align:middle line:90% OK, all right. 00:56:48.570 --> 00:56:49.862 align:middle line:90% So I'm not just making this up. 00:56:49.862 --> 00:56:51.067 align:middle line:90% That's good to know. 00:56:51.067 --> 00:56:52.650 align:middle line:84% I just have to check for these things. 00:56:52.650 --> 00:56:56.400 align:middle line:84% There are many, many idioms that mean "die." 00:56:56.400 --> 00:56:59.040 align:middle line:84% "Kick the bucket" and "buy the farm" are two of them. 00:56:59.040 --> 00:57:01.950 align:middle line:84% "Spill the beans," just to get things slightly less violent, 00:57:01.950 --> 00:57:04.800 align:middle line:84% spill the beans" means to reveal a secret. 00:57:04.800 --> 00:57:07.423 align:middle line:84% "Yawn in Technicolor," is that an idiom you guys know? 00:57:07.423 --> 00:57:08.340 align:middle line:90% "Yawn in Technicolor?" 00:57:08.340 --> 00:57:08.940 align:middle line:90% No? 00:57:08.940 --> 00:57:12.420 align:middle line:84% I got this one from a book of UCLA slang. 00:57:12.420 --> 00:57:13.587 align:middle line:90% There was a period there-- 00:57:13.587 --> 00:57:14.670 align:middle line:90% I hope they still do this. 00:57:14.670 --> 00:57:15.870 align:middle line:90% I'm not sure they do. 00:57:15.870 --> 00:57:18.623 align:middle line:84% The UCLA linguistics department had an undergrad class 00:57:18.623 --> 00:57:20.040 align:middle line:84% where the undergrads would compile 00:57:20.040 --> 00:57:24.750 align:middle line:84% a dictionary of undergrad slang, and they sold the dictionary. 00:57:24.750 --> 00:57:27.640 align:middle line:84% And you could buy the dictionary in the university bookstore. 00:57:27.640 --> 00:57:30.240 align:middle line:90% I bought a copy. 00:57:30.240 --> 00:57:33.060 align:middle line:84% There were many, many-- the UCLA Slang Dictionary 00:57:33.060 --> 00:57:34.200 align:middle line:90% is an interesting document. 00:57:34.200 --> 00:57:36.270 align:middle line:84% There are lots and lots of idioms that have 00:57:36.270 --> 00:57:41.160 align:middle line:90% to do with being drunk or high. 00:57:41.160 --> 00:57:43.110 align:middle line:84% All the idioms, the MIT idioms that I 00:57:43.110 --> 00:57:45.450 align:middle line:84% have to do with having too much work to do. 00:57:45.450 --> 00:57:47.080 align:middle line:90% [LAUGHTER] 00:57:47.080 --> 00:57:50.730 align:middle line:84% So I don't know if that's true, but I get that feeling. 00:57:50.730 --> 00:57:52.420 align:middle line:84% So yeah, the MIT slang dictionary, 00:57:52.420 --> 00:57:54.420 align:middle line:84% if there were an undergrad class like that here, 00:57:54.420 --> 00:57:56.760 align:middle line:84% I think it would be a less entertaining read. 00:57:56.760 --> 00:57:58.140 align:middle line:90% But I could be wrong. 00:57:58.140 --> 00:57:59.880 align:middle line:84% Anyway, that's where I learned this one. 00:57:59.880 --> 00:58:05.885 align:middle line:84% I once talked about this idiom in front of some Australians, 00:58:05.885 --> 00:58:07.260 align:middle line:84% and they were all like, oh, yeah, 00:58:07.260 --> 00:58:09.203 align:middle line:90% that's an Australian idiom. 00:58:09.203 --> 00:58:10.620 align:middle line:84% So it's possible that that's where 00:58:10.620 --> 00:58:13.310 align:middle line:84% it came to UCLA, that there were Australian students who 00:58:13.310 --> 00:58:14.310 align:middle line:90% were talking about that. 00:58:14.310 --> 00:58:16.920 align:middle line:90% Anyway, "yawn in Technicolor"-- 00:58:16.920 --> 00:58:17.940 align:middle line:90% impress your friends. 00:58:17.940 --> 00:58:21.090 align:middle line:84% Add this to your vocabulary, if it ever comes up. 00:58:21.090 --> 00:58:22.690 align:middle line:90% There are many, many idioms. 00:58:22.690 --> 00:58:24.720 align:middle line:84% So yeah, people are clear on what idioms are? 00:58:24.720 --> 00:58:29.100 align:middle line:84% So these are strings of words or structures that have literal 00:58:29.100 --> 00:58:30.920 align:middle line:84% meanings, I guess-- "yawn in Technicolor," 00:58:30.920 --> 00:58:33.960 align:middle line:90% maybe not so much-- 00:58:33.960 --> 00:58:36.120 align:middle line:84% but also have these non-literal meanings. 00:58:36.120 --> 00:58:39.720 align:middle line:84% And the non-literal meanings are quite-- 00:58:39.720 --> 00:58:41.250 align:middle line:90% what? 00:58:41.250 --> 00:58:42.180 align:middle line:90% Fragile, I guess. 00:58:42.180 --> 00:58:44.640 align:middle line:90% So "kick the bucket" means die. 00:58:44.640 --> 00:58:48.870 align:middle line:84% It also has a literal meaning where you punt a pail. 00:58:48.870 --> 00:58:54.540 align:middle line:84% But you can't change "kick" or "bucket" with a synonym 00:58:54.540 --> 00:58:56.190 align:middle line:90% and still mean "die." 00:58:56.190 --> 00:58:59.513 align:middle line:84% So if he "punts the bucket," that doesn't mean he died. 00:58:59.513 --> 00:59:01.680 align:middle line:84% It just means he literally kicked the bucket, right? 00:59:01.680 --> 00:59:05.250 align:middle line:84% Or if he "kicks a pail," if he "kicked the pail," 00:59:05.250 --> 00:59:06.780 align:middle line:90% that doesn't mean he died. 00:59:06.780 --> 00:59:10.680 align:middle line:84% So it has to be these words to have 00:59:10.680 --> 00:59:13.680 align:middle line:90% it have its idiomatic meaning. 00:59:13.680 --> 00:59:14.790 align:middle line:90% Idioms are fun. 00:59:14.790 --> 00:59:18.000 align:middle line:84% If you get a chance, you can ask your language consultant 00:59:18.000 --> 00:59:19.620 align:middle line:84% whether they can think of any idioms 00:59:19.620 --> 00:59:23.310 align:middle line:90% that you can think about. 00:59:23.310 --> 00:59:24.460 align:middle line:90% Now, here's the thing. 00:59:24.460 --> 00:59:27.180 align:middle line:84% There are a zillion idioms like these 00:59:27.180 --> 00:59:30.630 align:middle line:84% where the idiom consists of a verb and some phrase 00:59:30.630 --> 00:59:32.070 align:middle line:90% that the verb combines with. 00:59:32.070 --> 00:59:34.745 align:middle line:84% So a verb and its object, like in the first three examples, 00:59:34.745 --> 00:59:36.120 align:middle line:84% or a verb in prepositional phrase 00:59:36.120 --> 00:59:37.230 align:middle line:90% like in the last example-- 00:59:37.230 --> 00:59:39.390 align:middle line:84% lots and lots of idioms like that. 00:59:39.390 --> 00:59:41.430 align:middle line:84% So to say that that's the idiom is 00:59:41.430 --> 00:59:43.840 align:middle line:84% to say that if you have this idiom in a sentence-- 00:59:43.840 --> 00:59:47.010 align:middle line:84% so if you say something like, "John kicked the bucket," 00:59:47.010 --> 00:59:48.720 align:middle line:84% the idiomatic part of that sentence 00:59:48.720 --> 00:59:49.980 align:middle line:90% is "kick the bucket," right? 00:59:49.980 --> 00:59:52.170 align:middle line:90% "John" is not part of the idiom. 00:59:52.170 --> 00:59:53.340 align:middle line:90% Anybody can kick the bucket. 00:59:53.340 --> 00:59:54.945 align:middle line:90% In fact, all of us will someday. 00:59:54.945 --> 00:59:56.880 align:middle line:90% [LAUGHTER] 00:59:56.880 --> 01:00:00.757 align:middle line:84% Sorry, not to be depressing or anything. 01:00:00.757 --> 01:00:02.340 align:middle line:84% So the subject can be anything at all. 01:00:02.340 --> 01:00:03.923 align:middle line:84% It's the verb and the object that make 01:00:03.923 --> 01:00:05.490 align:middle line:90% the whole thing an idiom, yeah? 01:00:05.490 --> 01:00:07.140 align:middle line:84% And then there's the non-idiomatic part 01:00:07.140 --> 01:00:09.550 align:middle line:84% that you add to the idiom to make a sentence. 01:00:09.550 --> 01:00:11.350 align:middle line:90% Does that make sense? 01:00:11.350 --> 01:00:14.347 align:middle line:84% So idioms are typically parts of sentences 01:00:14.347 --> 01:00:16.930 align:middle line:84% that combine with other things to make complete sentences, not 01:00:16.930 --> 01:00:19.130 align:middle line:90% other non-idiomatic stuff. 01:00:19.130 --> 01:00:22.180 align:middle line:84% But here's the thing-- there are restrictions on idioms. 01:00:22.180 --> 01:00:25.270 align:middle line:84% There are no idioms that consist of a subject and a verb 01:00:25.270 --> 01:00:27.800 align:middle line:84% where the object can be anything. 01:00:27.800 --> 01:00:30.130 align:middle line:84% So I'm just giving this as an example. 01:00:30.130 --> 01:00:32.990 align:middle line:84% It's always hard to exemplify this because, of course, 01:00:32.990 --> 01:00:35.200 align:middle line:84% most strings of words are not idioms. 01:00:35.200 --> 01:00:39.637 align:middle line:84% But the last slide showed you, there are a zillion idioms 01:00:39.637 --> 01:00:41.470 align:middle line:84% where the verb and the object are an idiom-- 01:00:41.470 --> 01:00:43.760 align:middle line:84% "kick the bucket," "buy the farm"-- 01:00:43.760 --> 01:00:46.820 align:middle line:84% but there are no idioms where it's the subject and the verb. 01:00:46.820 --> 01:00:50.290 align:middle line:84% So here's an example of a non-idiom. 01:00:50.290 --> 01:00:53.620 align:middle line:84% There isn't an idiom, "the armadillo bit (blank)." 01:00:53.620 --> 01:00:55.930 align:middle line:84% So you can imagine an idiom where 01:00:55.930 --> 01:00:58.810 align:middle line:84% if I said, "The armadillo bit John yesterday," 01:00:58.810 --> 01:01:03.260 align:middle line:84% it meant "John was very busy yesterday" or something. 01:01:03.260 --> 01:01:06.850 align:middle line:84% Or, "The armadillo bit me this morning" 01:01:06.850 --> 01:01:08.832 align:middle line:84% means I have a headache, or something. 01:01:08.832 --> 01:01:10.790 align:middle line:84% You can imagine it meaning something like that. 01:01:10.790 --> 01:01:12.010 align:middle line:90% But that is not an idiom. 01:01:12.010 --> 01:01:14.555 align:middle line:84% And more importantly, there are no idioms like that. 01:01:14.555 --> 01:01:15.055 align:middle line:90% Yes? 01:01:15.055 --> 01:01:18.430 align:middle line:84% AUDIENCE: What about like "The cat got your tongue"? 01:01:18.430 --> 01:01:21.700 align:middle line:84% NORVIN RICHARDS: Well, that's a nice example. 01:01:21.700 --> 01:01:24.340 align:middle line:84% Notice, though, that the object is part of the idiom, right? 01:01:24.340 --> 01:01:32.980 align:middle line:84% So actually, "Cat got (blank)'s tongue." 01:01:32.980 --> 01:01:36.130 align:middle line:84% Actually, can the possessor of the tongue be anyone at all? 01:01:36.130 --> 01:01:39.460 align:middle line:84% Can I say, "The cat got John's tongue?" 01:01:39.460 --> 01:01:42.490 align:middle line:84% Meaning John had a hard time thinking of what to say? 01:01:42.490 --> 01:01:44.260 align:middle line:84% I think the easiest way for me to use 01:01:44.260 --> 01:01:48.190 align:middle line:84% this idiom is to ask someone, "Cat got your tongue?" 01:01:48.190 --> 01:01:51.400 align:middle line:84% If I have a hard time coming up with something to say, 01:01:51.400 --> 01:01:54.540 align:middle line:84% can I say "The cat got my tongue?" 01:01:54.540 --> 01:01:56.250 align:middle line:84% So there may or may not be a blank here. 01:01:56.250 --> 01:01:58.080 align:middle line:90% Maybe there should be "or" here. 01:01:58.080 --> 01:02:00.600 align:middle line:84% But anyway, the subject is indeed part of the idiom, 01:02:00.600 --> 01:02:02.110 align:middle line:90% but so is the object. 01:02:02.110 --> 01:02:04.440 align:middle line:90% So this is a nice-- 01:02:04.440 --> 01:02:07.320 align:middle line:84% you're making me refine what I just said, which is important. 01:02:07.320 --> 01:02:09.520 align:middle line:84% It's not the idioms can't contain the subject. 01:02:09.520 --> 01:02:10.020 align:middle line:90% They can. 01:02:10.020 --> 01:02:11.250 align:middle line:90% This is a nice example. 01:02:11.250 --> 01:02:13.960 align:middle line:84% But they also have to contain the object. 01:02:13.960 --> 01:02:14.460 align:middle line:90% Yes? 01:02:14.460 --> 01:02:17.200 align:middle line:84% What if [INAUDIBLE] something like "his tongue." 01:02:17.200 --> 01:02:18.770 align:middle line:90% The cat got his tongue-- 01:02:18.770 --> 01:02:19.720 align:middle line:90% can you say that? 01:02:19.720 --> 01:02:23.020 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:02:23.020 --> 01:02:25.540 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:02:25.540 --> 01:02:26.280 align:middle line:90% AUDIENCE: Yeah. 01:02:26.280 --> 01:02:27.610 align:middle line:90% NORVIN RICHARDS: Yeah, my son-- 01:02:27.610 --> 01:02:29.943 align:middle line:84% AUDIENCE: That's when he's like, what's wrong with this? 01:02:29.943 --> 01:02:30.640 align:middle line:90% 01:02:30.640 --> 01:02:32.140 align:middle line:84% NORVIN RICHARDS: Cat got his tongue. 01:02:32.140 --> 01:02:35.680 align:middle line:84% My son, who's 11, is shy, and he's 01:02:35.680 --> 01:02:38.140 align:middle line:84% having a hard time saying anything in front 01:02:38.140 --> 01:02:39.430 align:middle line:90% of people he doesn't know. 01:02:39.430 --> 01:02:41.772 align:middle line:84% And I say, "Oh, the cat got his tongue." 01:02:41.772 --> 01:02:43.710 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:02:43.710 --> 01:02:46.110 align:middle line:84% NORVIN RICHARDS: Maybe, yeah, yeah, yeah. 01:02:46.110 --> 01:02:48.508 align:middle line:84% So and this is why I waffled about 01:02:48.508 --> 01:02:49.800 align:middle line:90% whether there was a blank here. 01:02:49.800 --> 01:02:52.180 align:middle line:84% The fact that there can be a blank here is interesting. 01:02:52.180 --> 01:02:53.075 align:middle line:90% But the point is-- 01:02:53.075 --> 01:02:54.450 align:middle line:84% there are several points to make. 01:02:54.450 --> 01:02:57.510 align:middle line:84% But one is, look, "tongue" is also part of this idiom. 01:02:57.510 --> 01:02:59.280 align:middle line:84% So the object is part of the idiom. 01:02:59.280 --> 01:03:02.490 align:middle line:84% You're absolutely right that you might have some options 01:03:02.490 --> 01:03:03.840 align:middle line:90% about what to put here. 01:03:03.840 --> 01:03:04.590 align:middle line:90% Raquel? 01:03:04.590 --> 01:03:08.510 align:middle line:84% AUDIENCE: I feel like if you just have a, quote, "idiom," 01:03:08.510 --> 01:03:09.435 align:middle line:90% it's just like a verb. 01:03:09.435 --> 01:03:10.310 align:middle line:90% NORVIN RICHARDS: Yep. 01:03:10.310 --> 01:03:11.390 align:middle line:84% AUDIENCE: And then, if you just say 01:03:11.390 --> 01:03:13.010 align:middle line:84% that that's an alternative meaning, 01:03:13.010 --> 01:03:16.820 align:middle line:84% or like a slang meaning of that verb, like "flame," like, 01:03:16.820 --> 01:03:17.630 align:middle line:90% get really angry-- 01:03:17.630 --> 01:03:17.990 align:middle line:90% NORVIN RICHARDS: Flame. 01:03:17.990 --> 01:03:18.860 align:middle line:84% AUDIENCE: --or something like that. 01:03:18.860 --> 01:03:20.510 align:middle line:84% Yeah, people don't call it an idiom. 01:03:20.510 --> 01:03:23.110 align:middle line:84% They just say, oh, yeah, "flame" is just [INAUDIBLE].. 01:03:23.110 --> 01:03:26.600 align:middle line:90% 01:03:26.600 --> 01:03:28.738 align:middle line:84% NORVIN RICHARDS: That's a nice example. 01:03:28.738 --> 01:03:30.530 align:middle line:84% So maybe this is a way to think about this. 01:03:30.530 --> 01:03:33.072 align:middle line:84% If you're thinking about a tree here-- we'll put a tree here. 01:03:33.072 --> 01:03:39.120 align:middle line:90% 01:03:39.120 --> 01:03:42.140 align:middle line:84% So what we're seeing is there are idioms-- 01:03:42.140 --> 01:03:43.913 align:middle line:84% I started by saying there idioms like this 01:03:43.913 --> 01:03:45.830 align:middle line:84% where this whole thing is the idiom-- the verb 01:03:45.830 --> 01:03:48.325 align:middle line:84% and a noun phrase after it like, "buy the farm." 01:03:48.325 --> 01:03:49.700 align:middle line:84% Raquel is raising the possibility 01:03:49.700 --> 01:03:51.367 align:middle line:84% that there could be idioms that are just 01:03:51.367 --> 01:03:53.600 align:middle line:84% this size, although, as you say, that's not usually 01:03:53.600 --> 01:03:54.830 align:middle line:90% what you call them. 01:03:54.830 --> 01:03:58.550 align:middle line:84% So maybe "flame," where the subject and the object 01:03:58.550 --> 01:03:59.690 align:middle line:90% are not part of that idiom. 01:03:59.690 --> 01:04:01.790 align:middle line:84% So I can flame you, you can flame me, 01:04:01.790 --> 01:04:02.840 align:middle line:90% anyone can flame anyone. 01:04:02.840 --> 01:04:03.950 align:middle line:90% Those are not part of the idiom. 01:04:03.950 --> 01:04:05.560 align:middle line:84% The idiom is just the verb, "flame." 01:04:05.560 --> 01:04:09.930 align:middle line:84% Yeah, so yeah, we could call that an idiom or not. 01:04:09.930 --> 01:04:13.612 align:middle line:84% And then, earlier pointed out, "cat got your tongue," 01:04:13.612 --> 01:04:15.320 align:middle line:84% there are bigger idioms where the subject 01:04:15.320 --> 01:04:17.570 align:middle line:84% and the verb and the object are all part of the idiom. 01:04:17.570 --> 01:04:19.670 align:middle line:84% And then, there's questions about how do you 01:04:19.670 --> 01:04:21.710 align:middle line:90% get a possessor into here? 01:04:21.710 --> 01:04:23.010 align:middle line:90% I'm trying to think about that. 01:04:23.010 --> 01:04:25.510 align:middle line:84% So a way to talk about this, then, is to say, yeah, idioms-- 01:04:25.510 --> 01:04:27.020 align:middle line:90% but the point that I'm making-- 01:04:27.020 --> 01:04:28.520 align:middle line:84% attempting to make-- is there aren't 01:04:28.520 --> 01:04:31.070 align:middle line:90% any idioms that are just this. 01:04:31.070 --> 01:04:33.320 align:middle line:84% I wish I had colored chalk or something. 01:04:33.320 --> 01:04:35.330 align:middle line:84% There aren't any idioms that are just the subject and the verb, 01:04:35.330 --> 01:04:36.930 align:middle line:84% and the object can be anything at all. 01:04:36.930 --> 01:04:39.020 align:middle line:84% There's no "The armadillo bit John." 01:04:39.020 --> 01:04:43.820 align:middle line:84% There's no "The kumquats have fooled Mary," right? 01:04:43.820 --> 01:04:46.280 align:middle line:84% Where "The kumquats have fooled Mary" 01:04:46.280 --> 01:04:50.120 align:middle line:90% means Mary is daydreaming. 01:04:50.120 --> 01:04:52.490 align:middle line:90% There aren't idioms like that. 01:04:52.490 --> 01:04:53.300 align:middle line:90% Yeah? 01:04:53.300 --> 01:04:56.195 align:middle line:84% AUDIENCE: So is it basically like it has to be [INAUDIBLE]?? 01:04:56.195 --> 01:04:58.070 align:middle line:84% NORVIN RICHARDS: It's looking like it, right? 01:04:58.070 --> 01:04:59.600 align:middle line:90% Yeah, something like that. 01:04:59.600 --> 01:05:03.680 align:middle line:84% Although, how you're going to get "his" or "your" here 01:05:03.680 --> 01:05:07.010 align:middle line:84% in "the cat got your tongue," that's an interesting question, 01:05:07.010 --> 01:05:09.980 align:middle line:84% because that's presumably part of this noun phrase. 01:05:09.980 --> 01:05:13.592 align:middle line:84% But sweeping that potential problem under the rug, 01:05:13.592 --> 01:05:15.050 align:middle line:84% yeah, it's just, [? those ?] idioms 01:05:15.050 --> 01:05:18.350 align:middle line:90% have to be constituents, yeah. 01:05:18.350 --> 01:05:20.120 align:middle line:90% Yeah? 01:05:20.120 --> 01:05:22.250 align:middle line:90% OK? 01:05:22.250 --> 01:05:24.680 align:middle line:84% I'm pausing here because here I am boldly 01:05:24.680 --> 01:05:26.160 align:middle line:84% asserting there is no-- of course, 01:05:26.160 --> 01:05:27.785 align:middle line:84% there is no idiom, "the armadillo bit." 01:05:27.785 --> 01:05:29.750 align:middle line:84% But if you pick any random verb and noun, 01:05:29.750 --> 01:05:32.060 align:middle line:90% it's probably not an idiom. 01:05:32.060 --> 01:05:35.600 align:middle line:84% So there also isn't an idiom, "bit the armadillo," right? 01:05:35.600 --> 01:05:37.745 align:middle line:84% That could have been an idiom, but it isn't. 01:05:37.745 --> 01:05:39.620 align:middle line:84% So "John bit the armadillo last night" 01:05:39.620 --> 01:05:42.440 align:middle line:84% that doesn't mean "John worked hard last night." 01:05:42.440 --> 01:05:44.990 align:middle line:84% Yeah, it could have, but it doesn't. 01:05:44.990 --> 01:05:47.820 align:middle line:84% So claims of this kind are always a little weird. 01:05:47.820 --> 01:05:50.815 align:middle line:84% You should all go home and try to think, annoy 01:05:50.815 --> 01:05:52.190 align:middle line:84% your friends, and your relatives, 01:05:52.190 --> 01:05:55.100 align:middle line:84% and your roommates trying to come up with idioms 01:05:55.100 --> 01:05:56.040 align:middle line:90% that have this shape. 01:05:56.040 --> 01:05:59.440 align:middle line:90% But there don't seem to be any. 01:05:59.440 --> 01:05:59.940 align:middle line:90% Yeah? 01:05:59.940 --> 01:06:00.390 align:middle line:90% OK, good. 01:06:00.390 --> 01:06:01.770 align:middle line:84% I'm going to move quickly to the next slide 01:06:01.770 --> 01:06:03.187 align:middle line:84% before somebody comes up with one. 01:06:03.187 --> 01:06:03.883 align:middle line:90% [LAUGHTER] 01:06:03.883 --> 01:06:06.300 align:middle line:84% This isn't because the subject can't be part of the idiom. 01:06:06.300 --> 01:06:07.200 align:middle line:90% This has already come up. 01:06:07.200 --> 01:06:09.000 align:middle line:84% So "The cat got your tongue," "The cat is out of the bag," 01:06:09.000 --> 01:06:11.620 align:middle line:84% "The shit will hit the fan," there are lots of idioms. 01:06:11.620 --> 01:06:14.085 align:middle line:84% So "The shit will hit the fan" has a literal meaning. 01:06:14.085 --> 01:06:14.790 align:middle line:90% [LAUGHTER] 01:06:14.790 --> 01:06:17.130 align:middle line:84% But it also has a figurative meaning, 01:06:17.130 --> 01:06:20.250 align:middle line:84% like the situation will get dramatically worse quickly, 01:06:20.250 --> 01:06:21.490 align:middle line:90% or something like that. 01:06:21.490 --> 01:06:22.650 align:middle line:90% That's what it means. 01:06:22.650 --> 01:06:24.317 align:middle line:84% "The cat is out of the bag," yeah, there 01:06:24.317 --> 01:06:26.010 align:middle line:84% could be a literal cat in a literal bag, 01:06:26.010 --> 01:06:28.020 align:middle line:84% or it could mean the secret is out. 01:06:28.020 --> 01:06:29.820 align:middle line:84% That's the other thing it can mean. 01:06:29.820 --> 01:06:32.280 align:middle line:84% There are many idioms about cats for some reason. 01:06:32.280 --> 01:06:35.850 align:middle line:90% 01:06:35.850 --> 01:06:38.290 align:middle line:84% So yeah, there are idioms like "John will buy the farm." 01:06:38.290 --> 01:06:40.373 align:middle line:84% There are idioms like "The shit will hit the fan." 01:06:40.373 --> 01:06:41.630 align:middle line:90% Yeah, that's what we've seen. 01:06:41.630 --> 01:06:44.130 align:middle line:84% And there are no idioms like "The armadillo will bite John." 01:06:44.130 --> 01:06:45.323 align:middle line:90% Good. 01:06:45.323 --> 01:06:47.365 align:middle line:84% Maybe it's as though idioms must be constituents. 01:06:47.365 --> 01:06:50.990 align:middle line:84% Yeah, nicely, nicely pointed out. 01:06:50.990 --> 01:06:55.320 align:middle line:84% So OK, good, now we know something about idioms. 01:06:55.320 --> 01:06:59.420 align:middle line:84% So "The shit hit the fan," it's a fine idiom. 01:06:59.420 --> 01:07:01.530 align:middle line:84% "The shit seemed to hit the fan," 01:07:01.530 --> 01:07:04.530 align:middle line:84% "The shit seemed to be likely to hit the fan," 01:07:04.530 --> 01:07:07.740 align:middle line:84% these are all fine, and they all have idiomatic readings. 01:07:07.740 --> 01:07:10.630 align:middle line:90% 01:07:10.630 --> 01:07:12.510 align:middle line:84% But the things that I have circled 01:07:12.510 --> 01:07:15.660 align:middle line:84% are surely not constituents, right? 01:07:15.660 --> 01:07:18.120 align:middle line:84% So "The shit hit the fan" is a constituent, 01:07:18.120 --> 01:07:20.160 align:middle line:84% but "The shit seemed to hit the fan"-- "seem to" 01:07:20.160 --> 01:07:21.450 align:middle line:90% is not part of the idiom. 01:07:21.450 --> 01:07:24.780 align:middle line:84% These sentences are OK, and they have their idiomatic meanings. 01:07:24.780 --> 01:07:28.770 align:middle line:84% "Everything was fine all day until suddenly, at 12:35, 01:07:28.770 --> 01:07:31.140 align:middle line:84% the shit seemed to hit the fan"-- fine. 01:07:31.140 --> 01:07:33.120 align:middle line:90% Sad, of course, but grammatical. 01:07:33.120 --> 01:07:35.670 align:middle line:90% 01:07:35.670 --> 01:07:38.550 align:middle line:84% But "seem to" is not part of the idiom. 01:07:38.550 --> 01:07:40.560 align:middle line:84% It's like John in "John bought the farm." 01:07:40.560 --> 01:07:43.170 align:middle line:84% The idiom is, "The shit hit the fan." 01:07:43.170 --> 01:07:45.330 align:middle line:84% Or "The shit hit the fan" in "The shit seemed 01:07:45.330 --> 01:07:46.860 align:middle line:90% to be likely to hit the fan." 01:07:46.860 --> 01:07:50.550 align:middle line:84% So "I walked into the office, and everything 01:07:50.550 --> 01:07:51.510 align:middle line:90% seemed very precarious. 01:07:51.510 --> 01:07:52.280 align:middle line:90% Tempers were high. 01:07:52.280 --> 01:07:54.780 align:middle line:84% The shit seemed to be likely to hit the fan at any minute"-- 01:07:54.780 --> 01:07:56.130 align:middle line:90% yeah, perfectly fine. 01:07:56.130 --> 01:07:57.870 align:middle line:90% Has the idiomatic reading. 01:07:57.870 --> 01:08:01.580 align:middle line:84% But these things are quite far apart from each other. 01:08:01.580 --> 01:08:04.000 align:middle line:84% But if we believe in NP-movement, 01:08:04.000 --> 01:08:05.650 align:middle line:90% then we have a story. 01:08:05.650 --> 01:08:07.010 align:middle line:90% The story goes like this. 01:08:07.010 --> 01:08:09.160 align:middle line:84% How do you say, "The shit seemed to hit the fan"? 01:08:09.160 --> 01:08:11.290 align:middle line:84% Well, yeah, "the shit" and "hit the fan" 01:08:11.290 --> 01:08:14.320 align:middle line:84% are far apart from each other when you're done, 01:08:14.320 --> 01:08:18.520 align:middle line:84% but you start off with "seemed, the shit to hit the fan," 01:08:18.520 --> 01:08:22.689 align:middle line:84% and then "the shit" moved into the specifier of TP 01:08:22.689 --> 01:08:24.010 align:middle line:90% in the higher clause. 01:08:24.010 --> 01:08:26.800 align:middle line:84% That subject moves up to become the subject of the higher 01:08:26.800 --> 01:08:28.210 align:middle line:90% thing because of the EPP. 01:08:28.210 --> 01:08:31.870 align:middle line:84% The EPP wants there to be a subject up there. 01:08:31.870 --> 01:08:34.450 align:middle line:84% Another option would be to put an expletive there and say, 01:08:34.450 --> 01:08:36.850 align:middle line:84% "It seemed that the shit had hit the fan," 01:08:36.850 --> 01:08:38.308 align:middle line:84% or whatever-- to put an "it" there. 01:08:38.308 --> 01:08:39.767 align:middle line:84% But if you're not going to do that, 01:08:39.767 --> 01:08:42.310 align:middle line:84% you need to move the subject of the lower clause up 01:08:42.310 --> 01:08:43.810 align:middle line:90% into the higher clause. 01:08:43.810 --> 01:08:46.420 align:middle line:84% The consequence of that is that by the time you're done, 01:08:46.420 --> 01:08:48.819 align:middle line:84% "The shit hit the fan" is no longer a constituent. 01:08:48.819 --> 01:08:52.779 align:middle line:84% But it was once, and apparently, that's what matters. 01:08:52.779 --> 01:08:54.642 align:middle line:84% It needs to start as a constituent. 01:08:54.642 --> 01:08:56.350 align:middle line:84% Maybe one way to think about it would be, 01:08:56.350 --> 01:09:00.370 align:middle line:84% the idiom is a really big and complicated lexical item. 01:09:00.370 --> 01:09:03.154 align:middle line:84% If you're going to merge it, you need to merge it all at once. 01:09:03.154 --> 01:09:04.779 align:middle line:84% It's one way to think about it, anyway. 01:09:04.779 --> 01:09:05.950 align:middle line:90% Yeah? 01:09:05.950 --> 01:09:07.115 align:middle line:90% AUDIENCE: Yeah, question. 01:09:07.115 --> 01:09:07.580 align:middle line:90% NORVIN RICHARDS: Yeah. 01:09:07.580 --> 01:09:09.122 align:middle line:84% AUDIENCE: So first of all, what would 01:09:09.122 --> 01:09:11.060 align:middle line:84% happen if you wanted to say, "The shit 01:09:11.060 --> 01:09:12.529 align:middle line:90% was going to hit the fan?" 01:09:12.529 --> 01:09:15.450 align:middle line:84% NORVIN RICHARDS: Oh, good question. 01:09:15.450 --> 01:09:20.989 align:middle line:90% So "The shit hit the fan." 01:09:20.989 --> 01:09:24.149 align:middle line:90% 01:09:24.149 --> 01:09:26.600 align:middle line:90% "The shit--" very nice point. 01:09:26.600 --> 01:09:27.770 align:middle line:90% "The shit will hit the fan." 01:09:27.770 --> 01:09:30.720 align:middle line:90% 01:09:30.720 --> 01:09:33.060 align:middle line:90% "The shit might hit the fan." 01:09:33.060 --> 01:09:37.380 align:middle line:84% "The shit is going to hit the fan" Yeah, there are lots 01:09:37.380 --> 01:09:38.630 align:middle line:90% of things you can put in here. 01:09:38.630 --> 01:09:42.109 align:middle line:90% 01:09:42.109 --> 01:09:44.880 align:middle line:84% There is-- that a very nice point. 01:09:44.880 --> 01:09:46.880 align:middle line:84% There is a classic response to this point, which 01:09:46.880 --> 01:09:50.090 align:middle line:84% is to say, yeah, what you've discovered is there's even more 01:09:50.090 --> 01:09:52.010 align:middle line:90% NP-movement than we thought. 01:09:52.010 --> 01:09:57.617 align:middle line:84% So if I-- let me get a fresh board down here. 01:09:57.617 --> 01:09:58.700 align:middle line:90% How are we doing for time? 01:09:58.700 --> 01:10:02.180 align:middle line:84% Yeah, this would be a good place to stop, probably. 01:10:02.180 --> 01:10:04.400 align:middle line:84% The way I have been drawing trees for you, 01:10:04.400 --> 01:10:18.840 align:middle line:84% I've been drawing trees like, "John will eat the kumquats." 01:10:18.840 --> 01:10:21.480 align:middle line:90% 01:10:21.480 --> 01:10:23.790 align:middle line:84% I've been drawing trees like this, yeah? 01:10:23.790 --> 01:10:24.870 align:middle line:90% Does this look familiar? 01:10:24.870 --> 01:10:26.870 align:middle line:84% Does anybody have any questions about this tree? 01:10:26.870 --> 01:10:29.400 align:middle line:84% This should be painless at this point. 01:10:29.400 --> 01:10:31.950 align:middle line:84% If it's painful, then you should stop me. 01:10:31.950 --> 01:10:33.420 align:middle line:90% Is this OK? 01:10:33.420 --> 01:10:35.457 align:middle line:84% So we've got a verb phrase that consists 01:10:35.457 --> 01:10:37.290 align:middle line:84% of a verb and an object, "eat the kumquats." 01:10:37.290 --> 01:10:39.280 align:middle line:84% "Eat" is selecting for the object. 01:10:39.280 --> 01:10:42.090 align:middle line:84% And then we've said that words like "will" 01:10:42.090 --> 01:10:49.148 align:middle line:84% are instances of T, tense, and that T is the phrase, whose 01:10:49.148 --> 01:10:50.940 align:middle line:84% specifier, the subjects it's in-- actually, 01:10:50.940 --> 01:10:53.160 align:middle line:84% we have a name now for the fact that the subject sits there. 01:10:53.160 --> 01:10:54.868 align:middle line:84% That's the Extended Projection Principle. 01:10:54.868 --> 01:10:55.750 align:middle line:90% AUDIENCE: Yeah. 01:10:55.750 --> 01:10:57.292 align:middle line:84% NORVIN RICHARDS: And we've said there 01:10:57.292 --> 01:11:00.930 align:middle line:84% is some force that demands that there be something here. 01:11:00.930 --> 01:11:03.660 align:middle line:84% That's part of why John was here. 01:11:03.660 --> 01:11:06.390 align:middle line:84% Your point, which is a very well-taken one, 01:11:06.390 --> 01:11:12.480 align:middle line:84% is that if I say, "The shit will hit the fan--" 01:11:12.480 --> 01:11:14.310 align:middle line:84% remind me to erase this before we leave-- 01:11:14.310 --> 01:11:23.410 align:middle line:90% 01:11:23.410 --> 01:11:26.560 align:middle line:84% I've been going quickly partly in the hope 01:11:26.560 --> 01:11:29.110 align:middle line:84% that nobody would notice this, but yes, problem. 01:11:29.110 --> 01:11:33.070 align:middle line:84% This and this seemed to be the constituent, the idiom, 01:11:33.070 --> 01:11:36.220 align:middle line:84% but "will" is not part of the idiom. 01:11:36.220 --> 01:11:38.860 align:middle line:84% And there's a classic way of dealing with this problem. 01:11:38.860 --> 01:11:40.943 align:middle line:84% This is one of the classic arguments for something 01:11:40.943 --> 01:11:42.820 align:middle line:84% that is now widely believed, which 01:11:42.820 --> 01:11:46.450 align:middle line:84% is that the subject, although it is in the specifier of TP, 01:11:46.450 --> 01:11:48.340 align:middle line:90% it doesn't actually start there. 01:11:48.340 --> 01:11:50.350 align:middle line:84% I've been drawing trees where it starts there, 01:11:50.350 --> 01:11:51.640 align:middle line:90% but it actually starts lower. 01:11:51.640 --> 01:11:53.350 align:middle line:84% It starts inside the verb phrase, 01:11:53.350 --> 01:11:54.790 align:middle line:90% and it moves up to here. 01:11:54.790 --> 01:11:56.054 align:middle line:90% AUDIENCE: [INAUDIBLE]? 01:11:56.054 --> 01:12:00.140 align:middle line:90% 01:12:00.140 --> 01:12:02.780 align:middle line:84% NORVIN RICHARDS: Yeah, so that the real derivation for this-- 01:12:02.780 --> 01:12:04.613 align:middle line:84% this is what I've been concealing from you-- 01:12:04.613 --> 01:12:12.320 align:middle line:84% but the real derivation from this starts with the subject 01:12:12.320 --> 01:12:13.430 align:middle line:90% inside the verb phrase-- 01:12:13.430 --> 01:12:16.010 align:middle line:84% I would say it's a specifier of the verb phrase-- 01:12:16.010 --> 01:12:20.240 align:middle line:84% and it is raising to the specifier of TP always. 01:12:20.240 --> 01:12:23.630 align:middle line:84% So I've been talking so far as though the specifier or TP-- 01:12:23.630 --> 01:12:25.040 align:middle line:84% we've talked now about some cases 01:12:25.040 --> 01:12:27.860 align:middle line:84% where something moves into the specifier of TP. 01:12:27.860 --> 01:12:31.760 align:middle line:84% But from the earlier kinds of trees I was drawing for you, 01:12:31.760 --> 01:12:34.250 align:middle line:84% I was making it sound like the specifier of TP, sometimes, 01:12:34.250 --> 01:12:35.542 align:middle line:90% something is just merged there. 01:12:35.542 --> 01:12:36.950 align:middle line:90% It doesn't move there. 01:12:36.950 --> 01:12:39.347 align:middle line:84% For exactly the reason that you've just pointed out, 01:12:39.347 --> 01:12:41.930 align:middle line:84% people think that that's wrong-- that really, what's happening 01:12:41.930 --> 01:12:44.750 align:middle line:84% is that the subject always starts lower than TP 01:12:44.750 --> 01:12:46.857 align:middle line:84% and raises into the specifier of TP. 01:12:46.857 --> 01:12:49.190 align:middle line:84% And in English, it has to raise into the specifier of TP 01:12:49.190 --> 01:12:50.870 align:middle line:90% because of the EPP. 01:12:50.870 --> 01:12:53.180 align:middle line:84% And this argument from the behavior of idioms, 01:12:53.180 --> 01:12:56.040 align:middle line:84% this is one of the classic arguments for that conclusion, 01:12:56.040 --> 01:12:56.540 align:middle line:90% yeah. 01:12:56.540 --> 01:12:59.120 align:middle line:84% This is something that was discovered 01:12:59.120 --> 01:13:03.080 align:middle line:84% in the late '80s people first started saying that. 01:13:03.080 --> 01:13:09.770 align:middle line:84% It's called the VP internal subject hypothesis. 01:13:09.770 --> 01:13:13.190 align:middle line:84% And by now, it's very widely assumed to be true. 01:13:13.190 --> 01:13:15.890 align:middle line:84% Yeah, I was planning to conceal this from you because this is 01:13:15.890 --> 01:13:19.190 align:middle line:84% 24.900, but I should know better than to try to conceal things 01:13:19.190 --> 01:13:19.790 align:middle line:90% from you guys. 01:13:19.790 --> 01:13:21.410 align:middle line:84% Yeah, does that answer your question? 01:13:21.410 --> 01:13:22.035 align:middle line:90% AUDIENCE: Yeah. 01:13:22.035 --> 01:13:23.660 align:middle line:90% NORVIN RICHARDS: Cool. 01:13:23.660 --> 01:13:26.630 align:middle line:84% AUDIENCE: Is this an example about knowledge theory, where 01:13:26.630 --> 01:13:28.990 align:middle line:84% it's like we wanted idiom to be a constituent, 01:13:28.990 --> 01:13:32.060 align:middle line:84% but it matters more to satisfy the EPP? 01:13:32.060 --> 01:13:33.680 align:middle line:90% NORVIN RICHARDS: No. 01:13:33.680 --> 01:13:35.035 align:middle line:90% I see what you mean. 01:13:35.035 --> 01:13:37.160 align:middle line:84% The way people standardly think about this, anyway, 01:13:37.160 --> 01:13:38.350 align:middle line:90% it's something more like-- 01:13:38.350 --> 01:13:42.350 align:middle line:90% 01:13:42.350 --> 01:13:44.460 align:middle line:84% you're raising an interesting question. 01:13:44.460 --> 01:13:46.910 align:middle line:84% It would be interesting to think of ways of distinguishing 01:13:46.910 --> 01:13:50.030 align:middle line:90% these theories from each other. 01:13:50.030 --> 01:13:53.750 align:middle line:84% The way people think about this, it's something more like-- 01:13:53.750 --> 01:13:56.230 align:middle line:90% it's sort of-- mm. 01:13:56.230 --> 01:13:58.020 align:middle line:84% Remember the Projection Principle? 01:13:58.020 --> 01:14:00.777 align:middle line:84% So in the Projection Principle, it says the verb absolutely, 01:14:00.777 --> 01:14:03.360 align:middle line:84% absolutely has to have a sister if it's transitive, let's say. 01:14:03.360 --> 01:14:05.490 align:middle line:84% It has to have a noun phrase as its sister. 01:14:05.490 --> 01:14:09.390 align:middle line:84% And then, there are these other forces, things like wh-movement 01:14:09.390 --> 01:14:11.250 align:middle line:84% that say, oh, but wait, I want this thing 01:14:11.250 --> 01:14:12.670 align:middle line:90% to be over here instead. 01:14:12.670 --> 01:14:17.440 align:middle line:84% And so we say, yeah, it starts here, and then it moves. 01:14:17.440 --> 01:14:20.370 align:middle line:84% So it's as though there are these mutually incompatible 01:14:20.370 --> 01:14:23.647 align:middle line:84% requirements which are met by having a derivation, 01:14:23.647 --> 01:14:25.230 align:middle line:84% where something starts in one position 01:14:25.230 --> 01:14:26.970 align:middle line:84% and then moves to another position. 01:14:26.970 --> 01:14:29.100 align:middle line:84% And the idea is that the Projection Principle 01:14:29.100 --> 01:14:32.250 align:middle line:84% is satisfied because, well, you satisfied it first, 01:14:32.250 --> 01:14:36.840 align:middle line:84% at the beginning, and then later you did other things. 01:14:36.840 --> 01:14:39.000 align:middle line:90% We're going to talk more-- 01:14:39.000 --> 01:14:41.460 align:middle line:84% I think I promised that we would only do another day or two 01:14:41.460 --> 01:14:45.435 align:middle line:84% of syntax, so I'm running out of room 01:14:45.435 --> 01:14:47.560 align:middle line:84% in which to promise that we will talk about things. 01:14:47.560 --> 01:14:49.393 align:middle line:84% But I think we'll have a chance to talk more 01:14:49.393 --> 01:14:51.198 align:middle line:84% about the nature of the derivation 01:14:51.198 --> 01:14:52.740 align:middle line:84% and the order in which you do things, 01:14:52.740 --> 01:14:54.990 align:middle line:84% like the order in which you have to satisfy things. 01:14:54.990 --> 01:14:57.365 align:middle line:84% But the standard way people talk about this kind of thing 01:14:57.365 --> 01:14:59.370 align:middle line:84% is more like the rule-ordering kinds of things 01:14:59.370 --> 01:15:00.037 align:middle line:90% we talked about. 01:15:00.037 --> 01:15:02.748 align:middle line:84% It's like there is an early stage at which you must obey 01:15:02.748 --> 01:15:04.290 align:middle line:84% the Projection Principle, or at which 01:15:04.290 --> 01:15:06.390 align:middle line:84% you must obey this condition on idioms, 01:15:06.390 --> 01:15:08.970 align:middle line:84% and then there are other things that come along and mess 01:15:08.970 --> 01:15:10.530 align:middle line:90% everything up subsequently. 01:15:10.530 --> 01:15:14.050 align:middle line:84% That's a way people talk about this. 01:15:14.050 --> 01:15:20.260 align:middle line:84% Any questions about idioms, about constituents? 01:15:20.260 --> 01:15:21.335 align:middle line:90% Anything else? 01:15:21.335 --> 01:15:22.580 align:middle line:90% NP-movement? 01:15:22.580 --> 01:15:23.830 align:middle line:90% Awesome. 01:15:23.830 --> 01:15:26.440 align:middle line:90% So let me make sure-- 01:15:26.440 --> 01:15:28.665 align:middle line:84% yes, this is an excellent place to stop. 01:15:28.665 --> 01:15:30.790 align:middle line:84% So let me ask again, are there any other questions? 01:15:30.790 --> 01:15:33.350 align:middle line:90% Or shall we just stop? 01:15:33.350 --> 01:15:34.490 align:middle line:90% Cool, good. 01:15:34.490 --> 01:15:36.980 align:middle line:84% Now you know slightly more than I was planning to teach you 01:15:36.980 --> 01:15:38.750 align:middle line:90% about the behavior of subjects. 01:15:38.750 --> 01:15:42.070 align:middle line:84% And we'll pick it up here next time. 01:15:42.070 --> 01:15:46.000 align:middle line:90%