WEBVTT 00:00:00.000 --> 00:00:01.928 align:middle line:90% [SQUEAKING] 00:00:01.928 --> 00:00:04.338 align:middle line:90% [RUSTLING] 00:00:04.338 --> 00:00:07.712 align:middle line:90% [CLICKING] 00:00:07.712 --> 00:00:10.998 align:middle line:90% 00:00:10.998 --> 00:00:13.290 align:middle line:84% NORVIN RICHARDS: So last time, we started doing syntax, 00:00:13.290 --> 00:00:18.240 align:middle line:84% and we had started drawing trees like this one. 00:00:18.240 --> 00:00:21.240 align:middle line:90% What I said, I think, was-- 00:00:21.240 --> 00:00:22.823 align:middle line:84% if you remember when we were looking-- 00:00:22.823 --> 00:00:25.240 align:middle line:84% when we were doing morphology and we were looking at words 00:00:25.240 --> 00:00:27.570 align:middle line:84% like "unlockable," we were convincing ourselves that it 00:00:27.570 --> 00:00:32.490 align:middle line:84% was profitable to think of words like that as not just 00:00:32.490 --> 00:00:35.250 align:middle line:90% consisting of three morphemes-- 00:00:35.250 --> 00:00:37.722 align:middle line:90% "UN-LOCK-ABLE." 00:00:37.722 --> 00:00:39.180 align:middle line:84% It does consist of three morphemes, 00:00:39.180 --> 00:00:41.640 align:middle line:84% but that it's useful to think of these three morphemes 00:00:41.640 --> 00:00:45.310 align:middle line:90% as being composed in pairs. 00:00:45.310 --> 00:00:48.690 align:middle line:84% So there are two ways to make the word "unlockable." 00:00:48.690 --> 00:00:50.400 align:middle line:84% You can put "un-" together with "lock," 00:00:50.400 --> 00:00:53.025 align:middle line:84% and then you can plop the result of that together with "-able," 00:00:53.025 --> 00:00:55.023 align:middle line:84% or you can put "lock" together with "-able," 00:00:55.023 --> 00:00:56.940 align:middle line:84% and put the result of that together with "un-" 00:00:56.940 --> 00:00:59.760 align:middle line:90% and those mean different things. 00:00:59.760 --> 00:01:01.530 align:middle line:84% That was what we were saying when we were 00:01:01.530 --> 00:01:03.330 align:middle line:90% talking about "unlockable." 00:01:03.330 --> 00:01:07.770 align:middle line:84% We're going to talk about syntax the same way. 00:01:07.770 --> 00:01:09.960 align:middle line:90% Yes, this is a string of words-- 00:01:09.960 --> 00:01:12.270 align:middle line:90% however many words it is-- 00:01:12.270 --> 00:01:15.930 align:middle line:84% 8, I guess-- but there's more to say. 00:01:15.930 --> 00:01:17.680 align:middle line:84% It isn't just a string of eight words. 00:01:17.680 --> 00:01:19.770 align:middle line:84% It's a string of 8 words that are composed 00:01:19.770 --> 00:01:22.437 align:middle line:84% via this pairwise operation of merge that 00:01:22.437 --> 00:01:24.270 align:middle line:84% takes pairs of things and puts them together 00:01:24.270 --> 00:01:27.008 align:middle line:90% to form larger things. 00:01:27.008 --> 00:01:28.050 align:middle line:90% The other thing we said-- 00:01:28.050 --> 00:01:32.760 align:middle line:84% so I've just gotten started on building a syntactic structure 00:01:32.760 --> 00:01:34.510 align:middle line:84% for this sentence here by saying, 00:01:34.510 --> 00:01:37.050 align:middle line:84% yeah, we've got these two nouns, "book" 00:01:37.050 --> 00:01:40.750 align:middle line:84% and "garage," which are preceded by determiner 00:01:40.750 --> 00:01:43.500 align:middle line:84% "the," which should be merged together with these nouns 00:01:43.500 --> 00:01:46.200 align:middle line:90% to form larger objects. 00:01:46.200 --> 00:01:52.650 align:middle line:84% In the case of "unlockable," we were saying affixes like "un-" 00:01:52.650 --> 00:01:56.280 align:middle line:84% are going to come with instructions like "Put me 00:01:56.280 --> 00:01:59.530 align:middle line:84% together with a verb and I'll give you a verb." 00:01:59.530 --> 00:02:03.218 align:middle line:84% And instructions-- sorry, morphemes like "-able" 00:02:03.218 --> 00:02:05.760 align:middle line:84% are going to come together with instructions that say "Put me 00:02:05.760 --> 00:02:08.740 align:middle line:84% together with a verb and I will give you an adjective." 00:02:08.740 --> 00:02:14.370 align:middle line:84% And so when we put two things together, we're adding these-- 00:02:14.370 --> 00:02:19.140 align:middle line:84% putting labels on these nodes that we create via merge. 00:02:19.140 --> 00:02:21.570 align:middle line:84% In the case of morphology, what label 00:02:21.570 --> 00:02:26.460 align:middle line:84% you put on the larger node is determined by the affixes. 00:02:26.460 --> 00:02:29.190 align:middle line:84% And we decided affixes have to come with instructions, 00:02:29.190 --> 00:02:31.650 align:middle line:84% like "I turn verbs into adjectives," 00:02:31.650 --> 00:02:34.800 align:middle line:84% or "I attach to adjectives, and the result is an adjective"-- 00:02:34.800 --> 00:02:36.330 align:middle line:90% things like that. 00:02:36.330 --> 00:02:39.630 align:middle line:84% For syntax, similarly, we're going 00:02:39.630 --> 00:02:42.105 align:middle line:84% to take these pairs of things and we're 00:02:42.105 --> 00:02:43.980 align:middle line:84% going to merge them together, and we're going 00:02:43.980 --> 00:02:46.710 align:middle line:90% to give a label to the result. 00:02:46.710 --> 00:02:48.570 align:middle line:84% In syntax, the label of the result 00:02:48.570 --> 00:02:51.030 align:middle line:84% is always going to be the label one of the two things 00:02:51.030 --> 00:02:52.540 align:middle line:90% that you put together. 00:02:52.540 --> 00:02:55.530 align:middle line:84% So in this case here where I merged "the" 00:02:55.530 --> 00:02:59.100 align:middle line:84% and "book" or I merged "the" and "garage," I'm 00:02:59.100 --> 00:03:01.500 align:middle line:84% going to give the result of that the label 00:03:01.500 --> 00:03:05.718 align:middle line:84% "N," because we decided that the rules-- 00:03:05.718 --> 00:03:08.010 align:middle line:84% whatever they are-- for determining which things can go 00:03:08.010 --> 00:03:11.280 align:middle line:90% where seem to care about the N. 00:03:11.280 --> 00:03:14.010 align:middle line:84% That is, there are a variety of things-- 00:03:14.010 --> 00:03:17.040 align:middle line:90% take a sentence like this one-- 00:03:17.040 --> 00:03:20.230 align:middle line:84% that can go in the slot occupied here by, 00:03:20.230 --> 00:03:22.118 align:middle line:90% for example, "the book." 00:03:22.118 --> 00:03:24.660 align:middle line:84% And what they have in common is that they have nouns in them. 00:03:24.660 --> 00:03:26.910 align:middle line:84% So you can say "I will find the book in the garage," 00:03:26.910 --> 00:03:29.820 align:middle line:84% you can say "I will find books in the garage," you can say 00:03:29.820 --> 00:03:32.070 align:middle line:84% "I will find purple books in the garage," 00:03:32.070 --> 00:03:34.505 align:middle line:84% you can say "I will find the purple books about syntax 00:03:34.505 --> 00:03:35.130 align:middle line:90% in the garage." 00:03:35.130 --> 00:03:36.300 align:middle line:84% There are various kinds of modifiers 00:03:36.300 --> 00:03:37.675 align:middle line:84% and other things that you can add 00:03:37.675 --> 00:03:39.360 align:middle line:84% or not, but what they all have in common 00:03:39.360 --> 00:03:41.050 align:middle line:90% is that there's a noun. 00:03:41.050 --> 00:03:44.160 align:middle line:84% And so we're going to reflect that fact 00:03:44.160 --> 00:03:46.650 align:middle line:90% by naming that larger object-- 00:03:46.650 --> 00:03:49.200 align:middle line:84% the result of merging "the" with book-- 00:03:49.200 --> 00:03:53.130 align:middle line:84% giving it the label "N" as well, passing on the fact 00:03:53.130 --> 00:03:57.810 align:middle line:84% that there's a noun in here to that node. 00:03:57.810 --> 00:04:00.840 align:middle line:84% In the case of morphology, like I said, 00:04:00.840 --> 00:04:04.680 align:middle line:84% we had nice clear rules for what label 00:04:04.680 --> 00:04:07.060 align:middle line:90% to put on each of these nodes. 00:04:07.060 --> 00:04:10.440 align:middle line:84% The rule was look at the affix and the affix will tell you 00:04:10.440 --> 00:04:11.440 align:middle line:90% what you should do. 00:04:11.440 --> 00:04:14.940 align:middle line:84% So the affixes say things like, if you merge me with a verb, 00:04:14.940 --> 00:04:17.880 align:middle line:84% the result will be a verb, or if you merge me with a verb, 00:04:17.880 --> 00:04:20.700 align:middle line:90% the result will be an adjective. 00:04:20.700 --> 00:04:22.860 align:middle line:84% Syntacticians have not gotten that far 00:04:22.860 --> 00:04:26.050 align:middle line:84% with figuring out the rules for how to label things, 00:04:26.050 --> 00:04:28.050 align:middle line:84% so I am just going to label things. 00:04:28.050 --> 00:04:31.620 align:middle line:84% And this is a current topic of fairly hot debate, actually, 00:04:31.620 --> 00:04:33.330 align:middle line:84% among syntacticians-- like, why are we 00:04:33.330 --> 00:04:34.230 align:middle line:90% getting the labels that we are? 00:04:34.230 --> 00:04:36.188 align:middle line:84% We're pretty clear on what the label should be, 00:04:36.188 --> 00:04:39.490 align:middle line:90% but why is not as clear. 00:04:39.490 --> 00:04:43.080 align:middle line:90% So that was one part of the-- 00:04:43.080 --> 00:04:46.520 align:middle line:84% start of building a tree for this sentence. 00:04:46.520 --> 00:04:48.040 align:middle line:84% Are there any questions about this? 00:04:48.040 --> 00:04:51.960 align:middle line:90% This is all quasi-review. 00:04:51.960 --> 00:04:56.160 align:middle line:84% So from here, we've created "the book" and "the garage." 00:04:56.160 --> 00:05:00.180 align:middle line:84% We decided "in the garage" is also a constituent. 00:05:00.180 --> 00:05:01.560 align:middle line:90% It deserves a node. 00:05:01.560 --> 00:05:04.230 align:middle line:84% It's a prepositional phrase, so we're going to merge 00:05:04.230 --> 00:05:09.540 align:middle line:84% "in" together with the node that we created by-- that first 00:05:09.540 --> 00:05:11.810 align:middle line:84% instance of merge, "the garage"-- that N-- 00:05:11.810 --> 00:05:14.990 align:middle line:84% that larger N. Yeah, so we'll merge P with that larger N 00:05:14.990 --> 00:05:19.850 align:middle line:84% and we'll get a larger P-- a phrase with the label P-- 00:05:19.850 --> 00:05:22.460 align:middle line:84% what you could call a prepositional phrase. 00:05:22.460 --> 00:05:25.700 align:middle line:84% And then we said similarly, "find"-- 00:05:25.700 --> 00:05:30.270 align:middle line:84% we're going to merge "find" together first with "the book," 00:05:30.270 --> 00:05:33.380 align:middle line:84% and then finally, merge those two things together 00:05:33.380 --> 00:05:36.080 align:middle line:84% to get this structure for that much of the sentence. 00:05:36.080 --> 00:05:37.665 align:middle line:90% And we'll stop here for now. 00:05:37.665 --> 00:05:40.290 align:middle line:84% We're going to build the rest of the sentence in just a second. 00:05:40.290 --> 00:05:41.810 align:middle line:84% And I think this was as far as we got. 00:05:41.810 --> 00:05:44.030 align:middle line:84% And I said, is anybody alarmed or disturbed by this. 00:05:44.030 --> 00:05:47.370 align:middle line:84% And several of you raised your hands, because, in fact, 00:05:47.370 --> 00:05:50.300 align:middle line:84% this is only one of a couple of possible structures this tree 00:05:50.300 --> 00:05:51.500 align:middle line:90% can have that's-- 00:05:51.500 --> 00:05:53.960 align:middle line:84% sorry, this is only one of a couple of possible structures 00:05:53.960 --> 00:05:55.860 align:middle line:90% this sentence can have. 00:05:55.860 --> 00:05:58.363 align:middle line:84% We're going to spend a lot of time talking about that today. 00:05:58.363 --> 00:06:00.780 align:middle line:84% You were one of the people who raised your hand last time. 00:06:00.780 --> 00:06:02.030 align:middle line:90% Yeah? 00:06:02.030 --> 00:06:04.155 align:middle line:84% AUDIENCE: So we have P for both N 00:06:04.155 --> 00:06:06.380 align:middle line:84% and raised in the variety [INAUDIBLE].. 00:06:06.380 --> 00:06:10.490 align:middle line:84% So is N by itself a prepositional phrase 00:06:10.490 --> 00:06:12.570 align:middle line:90% or is it just verb phrase? 00:06:12.570 --> 00:06:16.270 align:middle line:84% NORVIN RICHARDS: So good question. 00:06:16.270 --> 00:06:18.460 align:middle line:84% Let's see if I can answer your question 00:06:18.460 --> 00:06:23.020 align:middle line:84% in a way that will not cause more confusion than it 00:06:23.020 --> 00:06:25.990 align:middle line:90% gets rid of. 00:06:25.990 --> 00:06:28.930 align:middle line:84% I have used now the phrase "prepositional phrase" 00:06:28.930 --> 00:06:34.777 align:middle line:84% a couple of times, and "noun phrase," and "verb phrase," 00:06:34.777 --> 00:06:36.610 align:middle line:84% and I'm going to dramatically introduce what 00:06:36.610 --> 00:06:37.912 align:middle line:90% those mean in just a second. 00:06:37.912 --> 00:06:39.370 align:middle line:84% I think it's almost the next slide. 00:06:39.370 --> 00:06:41.370 align:middle line:84% Let me just ask you to hang on to that question. 00:06:41.370 --> 00:06:43.540 align:middle line:90% It's a good question. 00:06:43.540 --> 00:06:46.165 align:middle line:84% But are there other questions about this much-- this far? 00:06:46.165 --> 00:06:48.740 align:middle line:90% 00:06:48.740 --> 00:06:50.740 align:middle line:84% So if you're looking at this and thinking, wait, 00:06:50.740 --> 00:06:52.690 align:middle line:84% I think of it as having a different structure, that's 00:06:52.690 --> 00:06:53.648 align:middle line:90% a good thought to have. 00:06:53.648 --> 00:06:55.880 align:middle line:90% Hang on to that thought. 00:06:55.880 --> 00:06:57.523 align:middle line:84% Ah, good-- this is the answer to your-- 00:06:57.523 --> 00:06:58.940 align:middle line:84% I'm sorry, what's your name again? 00:06:58.940 --> 00:06:59.750 align:middle line:90% AUDIENCE: Kirai. 00:06:59.750 --> 00:07:00.500 align:middle line:90% NORVIN RICHARDS: Say it again? 00:07:00.500 --> 00:07:01.125 align:middle line:90% AUDIENCE: Kirai 00:07:01.125 --> 00:07:02.390 align:middle line:90% NORVIN RICHARDS: Kirai. 00:07:02.390 --> 00:07:06.710 align:middle line:84% Kirai's question-- so he wanted to know, what do you 00:07:06.710 --> 00:07:08.390 align:middle line:90% mean when you say phrase? 00:07:08.390 --> 00:07:12.410 align:middle line:84% There's a convention, and it's really just 00:07:12.410 --> 00:07:16.520 align:middle line:84% to make it easier to look at these trees. 00:07:16.520 --> 00:07:19.633 align:middle line:84% But it's a very widespread convention, 00:07:19.633 --> 00:07:21.800 align:middle line:84% and I'll be doing it all the time when I draw trees, 00:07:21.800 --> 00:07:26.240 align:middle line:90% so let me introduce you to it. 00:07:26.240 --> 00:07:28.590 align:middle line:84% If you have a tree-- so the tree we had before-- 00:07:28.590 --> 00:07:30.200 align:middle line:90% let's go back a slide. 00:07:30.200 --> 00:07:32.360 align:middle line:90% The tree we had before-- whoa-- 00:07:32.360 --> 00:07:34.760 align:middle line:84% had various nodes that have the same label. 00:07:34.760 --> 00:07:37.460 align:middle line:84% There are three nodes here that have the label V, for example. 00:07:37.460 --> 00:07:40.340 align:middle line:84% And there are two nodes that have the label P, 00:07:40.340 --> 00:07:43.880 align:middle line:84% and there are four that have the label N, actually. 00:07:43.880 --> 00:07:45.470 align:middle line:84% One-- two associated with "garage" 00:07:45.470 --> 00:07:47.180 align:middle line:90% and two associated with "book." 00:07:47.180 --> 00:07:50.660 align:middle line:84% And if you look at this tree, it can be fairly confusing 00:07:50.660 --> 00:07:53.840 align:middle line:84% looking at all of these nodes that have identical labels. 00:07:53.840 --> 00:07:56.660 align:middle line:84% So the convention-- one convention 00:07:56.660 --> 00:07:59.090 align:middle line:84% just for making it easier to look at trees 00:07:59.090 --> 00:08:02.755 align:middle line:84% is to mark the highest node that has a given label when you have 00:08:02.755 --> 00:08:04.130 align:middle line:84% a sequence of nodes that all have 00:08:04.130 --> 00:08:07.220 align:middle line:84% the same label, because they all started with one word 00:08:07.220 --> 00:08:10.197 align:middle line:84% and they're inheriting a label from that word. 00:08:10.197 --> 00:08:12.030 align:middle line:84% When you get to the highest of those things, 00:08:12.030 --> 00:08:15.800 align:middle line:84% you give it the label P, where P stands for phrase. 00:08:15.800 --> 00:08:20.960 align:middle line:84% So the answer to this question-- when we say a prepositional 00:08:20.960 --> 00:08:24.590 align:middle line:84% phrase, what we typically mean is the largest thing with 00:08:24.590 --> 00:08:26.690 align:middle line:90% the label P from a given-- 00:08:26.690 --> 00:08:27.650 align:middle line:90% from a given node. 00:08:27.650 --> 00:08:30.230 align:middle line:84% So "in" by itself is a preposition, 00:08:30.230 --> 00:08:32.179 align:middle line:84% and that larger thing with the label P 00:08:32.179 --> 00:08:35.080 align:middle line:90% is the prepositional phrase. 00:08:35.080 --> 00:08:37.690 align:middle line:84% Again, this is just to make it easier to process trees 00:08:37.690 --> 00:08:39.429 align:middle line:84% so we won't see so many identical labels. 00:08:39.429 --> 00:08:41.110 align:middle line:84% We're just putting a mark saying, 00:08:41.110 --> 00:08:43.073 align:middle line:84% this is as far as this label got. 00:08:43.073 --> 00:08:43.990 align:middle line:90% That's all that means. 00:08:43.990 --> 00:08:45.212 align:middle line:90% Yeah? 00:08:45.212 --> 00:08:48.060 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:08:48.060 --> 00:08:51.097 align:middle line:90% 00:08:51.097 --> 00:08:51.930 align:middle line:90% NORVIN RICHARDS: Ah. 00:08:51.930 --> 00:08:54.810 align:middle line:84% So in this tree, it's not because there 00:08:54.810 --> 00:08:58.380 align:middle line:84% is another thing with the label V that is higher 00:08:58.380 --> 00:09:00.150 align:middle line:90% than "find the book"-- yes. 00:09:00.150 --> 00:09:02.820 align:middle line:84% But you are right-- there is another thing 00:09:02.820 --> 00:09:04.360 align:middle line:90% we do for exactly that case. 00:09:04.360 --> 00:09:05.640 align:middle line:90% Let me show you that. 00:09:05.640 --> 00:09:08.070 align:middle line:84% Nice question-- it allows me to segue nicely 00:09:08.070 --> 00:09:09.460 align:middle line:90% into the next slide. 00:09:09.460 --> 00:09:12.900 align:middle line:84% We have another thing that we do for exactly that case. 00:09:12.900 --> 00:09:15.242 align:middle line:90% Nodes that are not-- 00:09:15.242 --> 00:09:17.700 align:middle line:84% that don't get the label P, because they're not the highest 00:09:17.700 --> 00:09:20.520 align:middle line:90% thing and that also are not-- 00:09:20.520 --> 00:09:22.620 align:middle line:90% just the bear word, like find-- 00:09:22.620 --> 00:09:24.900 align:middle line:84% just the verb-- things that are in between that. 00:09:24.900 --> 00:09:27.720 align:middle line:90% It's an unfortunate label. 00:09:27.720 --> 00:09:34.200 align:middle line:84% We call things like-- so I'll draw it again here-- 00:09:34.200 --> 00:09:44.810 align:middle line:84% "find the book"-- so I won't draw the whole thing, 00:09:44.810 --> 00:09:46.818 align:middle line:84% but "find the book"-- there's this thing 00:09:46.818 --> 00:09:47.735 align:middle line:90% and then there's this. 00:09:47.735 --> 00:09:51.350 align:middle line:84% This is the VP because it's the highest thing with the label V. 00:09:51.350 --> 00:09:53.990 align:middle line:84% This is the V because it's just the verb-- it's just 00:09:53.990 --> 00:09:55.250 align:middle line:90% a word by itself. 00:09:55.250 --> 00:09:58.130 align:middle line:84% And this thing, we use a prime symbol to mark it, 00:09:58.130 --> 00:10:01.100 align:middle line:90% and it is called a V bar. 00:10:01.100 --> 00:10:05.810 align:middle line:84% I'm sorry, that's just the way we do things in syntax. 00:10:05.810 --> 00:10:07.340 align:middle line:84% I think it's called this because it 00:10:07.340 --> 00:10:11.360 align:middle line:84% used to be that people would write it like this 00:10:11.360 --> 00:10:13.520 align:middle line:90% with a bar over the top. 00:10:13.520 --> 00:10:15.290 align:middle line:90% This was in like the-- 00:10:15.290 --> 00:10:19.320 align:middle line:84% I don't know-- the 1960s, 1970s people began doing this. 00:10:19.320 --> 00:10:21.650 align:middle line:84% And I think people were literally using typewriters 00:10:21.650 --> 00:10:25.520 align:middle line:84% to go back and do an underscore over the top of the V 00:10:25.520 --> 00:10:26.540 align:middle line:90% or something like that. 00:10:26.540 --> 00:10:29.750 align:middle line:84% I think that's what they were doing with Stone Age word 00:10:29.750 --> 00:10:32.180 align:middle line:90% processing technology. 00:10:32.180 --> 00:10:35.090 align:middle line:84% So the result of all that is that although nobody writes 00:10:35.090 --> 00:10:37.820 align:middle line:84% V bar, typically, you will sometimes 00:10:37.820 --> 00:10:40.060 align:middle line:84% see this in textbooks and things like that. 00:10:40.060 --> 00:10:41.090 align:middle line:90% It's not very common. 00:10:41.090 --> 00:10:44.210 align:middle line:84% What people much more commonly do is just write V prime, 00:10:44.210 --> 00:10:46.430 align:middle line:84% but it is never ever called V prime. 00:10:46.430 --> 00:10:47.840 align:middle line:90% It's always called V bar. 00:10:47.840 --> 00:10:51.135 align:middle line:84% If you tell a syntactician, "Find me the V prime node," 00:10:51.135 --> 00:10:53.510 align:middle line:84% the syntactician will not know what you're talking about. 00:10:53.510 --> 00:10:54.590 align:middle line:90% Sorry. 00:10:54.590 --> 00:10:55.280 align:middle line:90% Yeah? 00:10:55.280 --> 00:10:58.565 align:middle line:84% AUDIENCE: So if "find the book" itself will not, like, 00:10:58.565 --> 00:11:00.545 align:middle line:84% considered a verb phrase, I guess, 00:11:00.545 --> 00:11:01.988 align:middle line:90% would it have the bar or? 00:11:01.988 --> 00:11:03.530 align:middle line:84% NORVIN RICHARDS: If "find the book"-- 00:11:03.530 --> 00:11:04.415 align:middle line:90% in this tree? 00:11:04.415 --> 00:11:05.040 align:middle line:90% AUDIENCE: Yeah. 00:11:05.040 --> 00:11:07.910 align:middle line:90% 00:11:07.910 --> 00:11:09.470 align:middle line:90% NORVIN RICHARDS: So it's a-- 00:11:09.470 --> 00:11:10.820 align:middle line:84% I'm sorry, can you ask your question again. 00:11:10.820 --> 00:11:12.528 align:middle line:84% AUDIENCE: Yeah, because I'm just thinking 00:11:12.528 --> 00:11:15.080 align:middle line:84% from grammatical standpoint, I guess, if "find the book" 00:11:15.080 --> 00:11:17.450 align:middle line:84% would be considered a verb phrase. 00:11:17.450 --> 00:11:20.010 align:middle line:84% But since it's not the highest standard of the bar, 00:11:20.010 --> 00:11:22.790 align:middle line:84% but if it were not a verb phrase, 00:11:22.790 --> 00:11:24.550 align:middle line:90% would it still not apply? 00:11:24.550 --> 00:11:26.460 align:middle line:84% NORVIN RICHARDS: So if this sentence 00:11:26.460 --> 00:11:28.560 align:middle line:84% ended with "find the book"-- if it were "I 00:11:28.560 --> 00:11:30.720 align:middle line:84% will find the book," then "find the book" 00:11:30.720 --> 00:11:34.035 align:middle line:90% would be a verb phrase. 00:11:34.035 --> 00:11:36.720 align:middle line:84% A phrase is just the name for the highest 00:11:36.720 --> 00:11:38.620 align:middle line:90% thing with that label. 00:11:38.620 --> 00:11:44.260 align:middle line:90% So just like in "the book"-- 00:11:44.260 --> 00:11:47.350 align:middle line:90% yeah, the book is a noun phrase. 00:11:47.350 --> 00:11:50.500 align:middle line:90% If you had something-- 00:11:50.500 --> 00:11:53.110 align:middle line:84% and "book" is not a noun phrase, it's just an N. 00:11:53.110 --> 00:11:55.720 align:middle line:84% So there are places where you don't 00:11:55.720 --> 00:11:57.520 align:middle line:84% get a bar just because you only have 00:11:57.520 --> 00:11:58.990 align:middle line:90% two things with that label. 00:11:58.990 --> 00:12:02.705 align:middle line:84% And you just get the N, and then there's the NP. 00:12:02.705 --> 00:12:04.330 align:middle line:84% In this particular case, you're getting 00:12:04.330 --> 00:12:07.990 align:middle line:84% a bar because there are three things with the label V. 00:12:07.990 --> 00:12:09.850 align:middle line:90% And if there were more things-- 00:12:09.850 --> 00:12:11.680 align:middle line:84% if this verb phrase were more complicated, 00:12:11.680 --> 00:12:13.480 align:middle line:84% you could end up with a bunch of bars 00:12:13.480 --> 00:12:15.190 align:middle line:90% in between the verb and the VP. 00:12:15.190 --> 00:12:17.860 align:middle line:84% In general, you're going to get one P up 00:12:17.860 --> 00:12:20.950 align:middle line:84% at the top and one thing with no extra doohickey 00:12:20.950 --> 00:12:24.150 align:middle line:84% down at the bottom, so just V. And everything in between it 00:12:24.150 --> 00:12:26.590 align:middle line:90% will be bar. 00:12:26.590 --> 00:12:28.090 align:middle line:84% Again, this is all just conventions. 00:12:28.090 --> 00:12:29.882 align:middle line:84% This is not meant to reflect anything deep. 00:12:29.882 --> 00:12:31.900 align:middle line:84% It's meant to make it possible for human beings 00:12:31.900 --> 00:12:33.580 align:middle line:84% to look at these trees and process 00:12:33.580 --> 00:12:34.690 align:middle line:90% the information in them. 00:12:34.690 --> 00:12:37.950 align:middle line:90% That's all it's for. 00:12:37.950 --> 00:12:39.210 align:middle line:90% Raquel? 00:12:39.210 --> 00:12:43.290 align:middle line:84% AUDIENCE: For things like "will find" or like an auxiliary 00:12:43.290 --> 00:12:45.155 align:middle line:84% where you squish multiple verbs together, 00:12:45.155 --> 00:12:47.280 align:middle line:84% is there something more complicated going on there, 00:12:47.280 --> 00:12:50.130 align:middle line:84% or could we say that they're both part of a verb phrase? 00:12:50.130 --> 00:12:51.870 align:middle line:84% NORVIN RICHARDS: We have not gotten there, but you're right. 00:12:51.870 --> 00:12:54.720 align:middle line:84% And so Raquel wants to know what am I going to do with "I will"? 00:12:54.720 --> 00:12:56.160 align:middle line:84% And I will-- I will tell you what 00:12:56.160 --> 00:12:57.760 align:middle line:90% I'm going to do with that now. 00:12:57.760 --> 00:13:01.220 align:middle line:84% We will give "will" its own label in just a second. 00:13:01.220 --> 00:13:02.730 align:middle line:90% Good question. 00:13:02.730 --> 00:13:05.680 align:middle line:90% Other questions about this? 00:13:05.680 --> 00:13:07.590 align:middle line:90% Reasonably clear? 00:13:07.590 --> 00:13:12.330 align:middle line:90% So-- but yeah. 00:13:12.330 --> 00:13:15.420 align:middle line:84% So again, P is the marker for the highest thing with a given 00:13:15.420 --> 00:13:16.458 align:middle line:90% label. 00:13:16.458 --> 00:13:18.750 align:middle line:84% The lowest thing with a given label doesn't get a mark. 00:13:18.750 --> 00:13:22.260 align:middle line:84% Sometimes, when people want to make it clear that it's not 00:13:22.260 --> 00:13:25.350 align:middle line:84% a bar and it's not a P, they'll put a little raised 0 on it. 00:13:25.350 --> 00:13:30.000 align:middle line:84% I won't do that, but you'll see people do that sometimes. 00:13:30.000 --> 00:13:32.650 align:middle line:84% So you've got the unmarked thing-- that's the word itself. 00:13:32.650 --> 00:13:34.483 align:middle line:84% You've got the P-- that's the highest thing, 00:13:34.483 --> 00:13:35.820 align:middle line:90% and everything else is a bar. 00:13:35.820 --> 00:13:37.320 align:middle line:84% One consequence of all this actually 00:13:37.320 --> 00:13:45.275 align:middle line:84% is that there are things which are both phrases and heads. 00:13:45.275 --> 00:13:48.510 align:middle line:84% This is connected to Kirai's question 00:13:48.510 --> 00:13:49.780 align:middle line:90% from a little while ago. 00:13:49.780 --> 00:13:54.870 align:middle line:84% So the D "the" in this-- both of the instances of D "the" 00:13:54.870 --> 00:13:58.320 align:middle line:84% in this tree, they are the lowest thing with the label D, 00:13:58.320 --> 00:13:59.970 align:middle line:90% so they don't get marked. 00:13:59.970 --> 00:14:02.550 align:middle line:84% They're also the highest thing with the label D, 00:14:02.550 --> 00:14:05.770 align:middle line:84% because neither of them is higher than the other. 00:14:05.770 --> 00:14:07.890 align:middle line:90% They're just two instances of D. 00:14:07.890 --> 00:14:12.330 align:middle line:84% And so an influential school of thought about what's 00:14:12.330 --> 00:14:13.930 align:middle line:84% going on in cases like that is to say, 00:14:13.930 --> 00:14:16.860 align:middle line:84% yeah, those are both phrases and not phrases. 00:14:16.860 --> 00:14:19.202 align:middle line:84% Try not to be too uptight about these labels. 00:14:19.202 --> 00:14:20.910 align:middle line:84% That's one way to talk about them anyway. 00:14:20.910 --> 00:14:24.480 align:middle line:90% 00:14:24.480 --> 00:14:31.340 align:middle line:90% So now, I drew you a tree-- 00:14:31.340 --> 00:14:34.772 align:middle line:84% and some of you last time objected. 00:14:34.772 --> 00:14:36.980 align:middle line:84% I drew you a tree in which a prepositional phrase was 00:14:36.980 --> 00:14:39.380 align:middle line:90% modifying a verb phrase. 00:14:39.380 --> 00:14:41.540 align:middle line:84% That's-- I'll show you the tree again. 00:14:41.540 --> 00:14:43.940 align:middle line:84% You've got a prepositional phrase, "in the garage." 00:14:43.940 --> 00:14:48.710 align:middle line:84% And I've got that prepositional phrase merged together 00:14:48.710 --> 00:14:50.600 align:middle line:84% with something with the label V. That's 00:14:50.600 --> 00:14:52.460 align:middle line:84% a prepositional phrase that's telling you 00:14:52.460 --> 00:14:54.668 align:middle line:84% where the finding of the book is going to take place. 00:14:54.668 --> 00:14:57.920 align:middle line:84% It's going to take place in the garage. 00:14:57.920 --> 00:15:00.680 align:middle line:84% Now, prepositional phrases can also modify noun phrases. 00:15:00.680 --> 00:15:02.810 align:middle line:84% You can say things like "I will find books 00:15:02.810 --> 00:15:05.497 align:middle line:84% about syntax," where "about syntax" is telling you 00:15:05.497 --> 00:15:07.205 align:middle line:84% what kinds of books you're going to find. 00:15:07.205 --> 00:15:10.430 align:middle line:90% 00:15:10.430 --> 00:15:12.500 align:middle line:84% Now, this is just some consequences 00:15:12.500 --> 00:15:15.590 align:middle line:90% of that we can think about. 00:15:15.590 --> 00:15:18.530 align:middle line:84% First, we've said that noun phrases can 00:15:18.530 --> 00:15:20.300 align:middle line:90% contain prepositional phrases-- 00:15:20.300 --> 00:15:22.587 align:middle line:90% books about syntax. 00:15:22.587 --> 00:15:24.170 align:middle line:84% And we know that prepositional phrases 00:15:24.170 --> 00:15:27.530 align:middle line:84% can contain noun phrases, which can 00:15:27.530 --> 00:15:31.530 align:middle line:84% contain prepositional phrases, which can contain noun phrases. 00:15:31.530 --> 00:15:33.950 align:middle line:84% So there is no reason to ever stop talking. 00:15:33.950 --> 00:15:36.450 align:middle line:84% We talked about this when we were talking about-- 00:15:36.450 --> 00:15:38.870 align:middle line:90% I'm wearing my mask again. 00:15:38.870 --> 00:15:42.085 align:middle line:84% We talked about this when we were talking about-- 00:15:42.085 --> 00:15:44.210 align:middle line:84% the competence / performance distinction-- the fact 00:15:44.210 --> 00:15:48.290 align:middle line:84% that we're idealizing the actual kinds of sentences 00:15:48.290 --> 00:15:49.055 align:middle line:90% that people say. 00:15:49.055 --> 00:15:52.650 align:middle line:90% 00:15:52.650 --> 00:15:55.320 align:middle line:90% Here's another case of this. 00:15:55.320 --> 00:15:57.540 align:middle line:84% So you can have a book about islands, 00:15:57.540 --> 00:15:59.670 align:middle line:84% and you can have a book about islands on lakes, 00:15:59.670 --> 00:16:02.040 align:middle line:84% you can have a book about islands on lakes on islands, 00:16:02.040 --> 00:16:03.790 align:middle line:84% you can have a book about islands on lakes 00:16:03.790 --> 00:16:05.190 align:middle line:90% on islands on lakes, and so on. 00:16:05.190 --> 00:16:07.410 align:middle line:84% Eventually, you will run out of things 00:16:07.410 --> 00:16:10.260 align:middle line:84% that your book can be about if you keep repeating this, 00:16:10.260 --> 00:16:11.865 align:middle line:84% but that is a fact about geography, 00:16:11.865 --> 00:16:14.700 align:middle line:84% and this is not a geography class. 00:16:14.700 --> 00:16:17.640 align:middle line:84% There actually is a website-- or used to be-- 00:16:17.640 --> 00:16:20.250 align:middle line:84% that listed the largest island on a lake, 00:16:20.250 --> 00:16:22.170 align:middle line:84% and the largest island on a lake on an island, 00:16:22.170 --> 00:16:24.540 align:middle line:84% and the largest island on a lake on an island on a lake. 00:16:24.540 --> 00:16:25.290 align:middle line:90% It's kind of cool. 00:16:25.290 --> 00:16:28.602 align:middle line:90% 00:16:28.602 --> 00:16:31.980 align:middle line:84% There's a lot of stuff going around Lake Taal 00:16:31.980 --> 00:16:34.860 align:middle line:90% in the Philippines I remember. 00:16:34.860 --> 00:16:35.520 align:middle line:90% Is this clear? 00:16:35.520 --> 00:16:36.450 align:middle line:90% So there's another-- 00:16:36.450 --> 00:16:37.890 align:middle line:84% I just said there is another instance where 00:16:37.890 --> 00:16:40.432 align:middle line:84% we might want to care about the difference between competence 00:16:40.432 --> 00:16:41.940 align:middle line:90% and performance. 00:16:41.940 --> 00:16:44.280 align:middle line:84% We've got a theory that we're starting 00:16:44.280 --> 00:16:48.930 align:middle line:84% to build about syntax that makes it possible for noun phrases 00:16:48.930 --> 00:16:51.450 align:middle line:90% to be arbitrarily long. 00:16:51.450 --> 00:16:55.290 align:middle line:84% No one's ever going to utter an arbitrarily long noun phrase. 00:16:55.290 --> 00:16:58.290 align:middle line:84% Nobody's going to keep talking as long as this grammar would 00:16:58.290 --> 00:17:00.570 align:middle line:84% allow you to, but that's a fact about, as I said, 00:17:00.570 --> 00:17:03.030 align:middle line:90% geography and life. 00:17:03.030 --> 00:17:06.119 align:middle line:84% People have better things to do than to continue 00:17:06.119 --> 00:17:08.200 align:middle line:90% repeating these things. 00:17:08.200 --> 00:17:12.450 align:middle line:84% But it's clear we want a grammar that allows this. 00:17:12.450 --> 00:17:15.000 align:middle line:84% So yeah, you couldn't actually utter an infinitely long noun 00:17:15.000 --> 00:17:16.667 align:middle line:84% phrase, but we now have a grammar that's 00:17:16.667 --> 00:17:19.780 align:middle line:90% capable of producing one. 00:17:19.780 --> 00:17:22.599 align:middle line:84% Another thing that it gets us-- and this is the heart 00:17:22.599 --> 00:17:25.930 align:middle line:84% of the reaction that I was getting last time when I showed 00:17:25.930 --> 00:17:28.900 align:middle line:84% you the tree that I wanted to do for that sentence-- 00:17:28.900 --> 00:17:30.830 align:middle line:84% is that there are cases of ambiguity. 00:17:30.830 --> 00:17:32.800 align:middle line:84% So this is a classic Marx brothers 00:17:32.800 --> 00:17:37.360 align:middle line:84% line in Animal Crackers, I think, 00:17:37.360 --> 00:17:41.110 align:middle line:84% Groucho Marx says to someone, "I once shot an elephant 00:17:41.110 --> 00:17:42.220 align:middle line:90% in my pajamas." 00:17:42.220 --> 00:17:44.440 align:middle line:84% And there's a pause, and he says, 00:17:44.440 --> 00:17:48.130 align:middle line:84% "How he got in my pajamas, I'll never know." 00:17:48.130 --> 00:17:52.098 align:middle line:84% So Groucho Marx here is making use of the ambiguity 00:17:52.098 --> 00:17:53.140 align:middle line:90% of the original sentence. 00:17:53.140 --> 00:17:56.590 align:middle line:84% "I shot an elephant in my pajamas" has two things 00:17:56.590 --> 00:17:59.230 align:middle line:84% that it could, in principle, mean. 00:17:59.230 --> 00:18:04.540 align:middle line:84% It could mean I shot an elephant while I was-- so here's 00:18:04.540 --> 00:18:05.245 align:middle line:90% in my pajamas. 00:18:05.245 --> 00:18:08.110 align:middle line:90% 00:18:08.110 --> 00:18:10.300 align:middle line:84% The prepositional phrase in my pajamas 00:18:10.300 --> 00:18:11.633 align:middle line:90% could modify the elephant. 00:18:11.633 --> 00:18:13.300 align:middle line:84% That is, it could be it was the elephant 00:18:13.300 --> 00:18:16.450 align:middle line:90% who was wearing the pajamas-- 00:18:16.450 --> 00:18:18.010 align:middle line:84% I shot an elephant in my pajamas. 00:18:18.010 --> 00:18:22.300 align:middle line:84% Or it could be that we put together "shoot an elephant," 00:18:22.300 --> 00:18:25.960 align:middle line:84% and then we merge "in my pajamas" with that 00:18:25.960 --> 00:18:29.782 align:middle line:84% so that "in my pajamas" is modifying the way in which I 00:18:29.782 --> 00:18:30.490 align:middle line:90% shot an elephant. 00:18:30.490 --> 00:18:33.315 align:middle line:84% That is, I shot an elephant while I was wearing my pajamas. 00:18:33.315 --> 00:18:35.180 align:middle line:84% And so I shot an elephant in my pajamas, 00:18:35.180 --> 00:18:36.350 align:middle line:90% it's an ambiguous sentence. 00:18:36.350 --> 00:18:39.070 align:middle line:84% It could mean either I was wearing pajamas 00:18:39.070 --> 00:18:41.890 align:middle line:84% or the elephant was wearing pajamas. 00:18:41.890 --> 00:18:44.980 align:middle line:90% Do people get that ambiguity? 00:18:44.980 --> 00:18:47.170 align:middle line:90% It's a classic joke. 00:18:47.170 --> 00:18:51.700 align:middle line:90% It's odd that we enjoy this. 00:18:51.700 --> 00:18:56.230 align:middle line:84% There's a lot of humor that has this shape, where someone says 00:18:56.230 --> 00:18:59.290 align:middle line:84% an ambiguous sentence that has one normal meaning and one 00:18:59.290 --> 00:19:02.140 align:middle line:84% strange meaning, and then the punchline 00:19:02.140 --> 00:19:04.630 align:middle line:84% reveals that they meant the strange meaning. 00:19:04.630 --> 00:19:05.860 align:middle line:90% That's the form of this joke. 00:19:05.860 --> 00:19:07.360 align:middle line:84% "I shot an elephant in my pajamas"-- 00:19:07.360 --> 00:19:08.680 align:middle line:84% you, of course, think that he means 00:19:08.680 --> 00:19:09.910 align:middle line:90% he was wearing his pajamas. 00:19:09.910 --> 00:19:11.410 align:middle line:84% And then he reveals that, no, it was 00:19:11.410 --> 00:19:13.160 align:middle line:84% the elephant that was wearing the pajamas, 00:19:13.160 --> 00:19:14.320 align:middle line:90% and then everybody laughs. 00:19:14.320 --> 00:19:15.730 align:middle line:90% As I said, it's odd that we-- 00:19:15.730 --> 00:19:19.210 align:middle line:84% I mean, normally, being told, haha, you misunderstood me, 00:19:19.210 --> 00:19:20.980 align:middle line:90% that's not fun. 00:19:20.980 --> 00:19:24.070 align:middle line:84% But we actually pay people to do this. 00:19:24.070 --> 00:19:24.970 align:middle line:90% It's kind of strange. 00:19:24.970 --> 00:19:25.930 align:middle line:90% Yeah? 00:19:25.930 --> 00:19:29.750 align:middle line:84% AUDIENCE: So it almost seems like the one on the left 00:19:29.750 --> 00:19:33.240 align:middle line:90% that's like started reading is-- 00:19:33.240 --> 00:19:34.657 align:middle line:84% almost seems like a simpler tree-- 00:19:34.657 --> 00:19:35.573 align:middle line:90% NORVIN RICHARDS: Uhuh. 00:19:35.573 --> 00:19:36.150 align:middle line:90% [LAUGHS] 00:19:36.150 --> 00:19:36.540 align:middle line:90% Yeah. 00:19:36.540 --> 00:19:38.748 align:middle line:84% AUDIENCE: --but compared to the more complicated one, 00:19:38.748 --> 00:19:41.292 align:middle line:84% is that some of the reason why this is ambiguous? 00:19:41.292 --> 00:19:42.750 align:middle line:84% The one that has the normal meaning 00:19:42.750 --> 00:19:44.850 align:middle line:84% has this complicated structure, but the one 00:19:44.850 --> 00:19:49.080 align:middle line:84% that has the wrong meaning is a very simple structure 00:19:49.080 --> 00:19:49.903 align:middle line:90% [INAUDIBLE]? 00:19:49.903 --> 00:19:50.820 align:middle line:90% NORVIN RICHARDS: Yeah. 00:19:50.820 --> 00:19:52.950 align:middle line:84% No, it's interesting that you have this-- 00:19:52.950 --> 00:19:55.290 align:middle line:84% I mean, right now, what we have is 00:19:55.290 --> 00:19:57.390 align:middle line:84% just a grammar that predicts that it could 00:19:57.390 --> 00:19:59.130 align:middle line:90% mean either of those things. 00:19:59.130 --> 00:20:01.950 align:middle line:84% So that prepositional phrase in my pajamas 00:20:01.950 --> 00:20:04.320 align:middle line:84% can be merged in a couple of different places, 00:20:04.320 --> 00:20:06.542 align:middle line:84% and the consequence is that it could mean 00:20:06.542 --> 00:20:07.750 align:middle line:90% a couple of different things. 00:20:07.750 --> 00:20:09.792 align:middle line:84% You're absolutely right that the tree on the left 00:20:09.792 --> 00:20:12.060 align:middle line:84% is prettier than the tree on the right. 00:20:12.060 --> 00:20:18.630 align:middle line:84% And people who work on how people deal with ambiguities 00:20:18.630 --> 00:20:21.380 align:middle line:84% like this in real-time-- this is something people work on-- 00:20:21.380 --> 00:20:26.850 align:middle line:84% processing-- develop theories about where the preferences are 00:20:26.850 --> 00:20:29.460 align:middle line:84% for where prepositional phrases should be attached 00:20:29.460 --> 00:20:32.843 align:middle line:84% in a place like this, whether there's a preference for trees 00:20:32.843 --> 00:20:34.260 align:middle line:84% like the tree on the left or trees 00:20:34.260 --> 00:20:35.430 align:middle line:90% like the tree on the right. 00:20:35.430 --> 00:20:36.960 align:middle line:84% I think there isn't supposed to be 00:20:36.960 --> 00:20:42.702 align:middle line:84% a general preference for trees like the tree on the left. 00:20:42.702 --> 00:20:44.160 align:middle line:84% I think this is just supposed to be 00:20:44.160 --> 00:20:47.880 align:middle line:84% a case where, in principle, it's ambiguous and-- 00:20:47.880 --> 00:20:49.553 align:middle line:84% but people who work on processing 00:20:49.553 --> 00:20:51.720 align:middle line:84% have found places where there's a preference for one 00:20:51.720 --> 00:20:53.970 align:middle line:84% kind of tree over another, so that is something people 00:20:53.970 --> 00:20:55.270 align:middle line:90% work on. 00:20:55.270 --> 00:20:56.536 align:middle line:90% Raquel? 00:20:56.536 --> 00:20:59.080 align:middle line:84% AUDIENCE: I guess a simple question-- 00:20:59.080 --> 00:21:02.780 align:middle line:84% so at the bottom of the left, there's an NP, like my pajamas, 00:21:02.780 --> 00:21:06.350 align:middle line:90% and there's N-bar that's higher. 00:21:06.350 --> 00:21:10.250 align:middle line:84% I guess why isn't an NP down there 00:21:10.250 --> 00:21:12.220 align:middle line:90% if there's another NP up above? 00:21:12.220 --> 00:21:14.500 align:middle line:84% NORVIN RICHARDS: Yeah, so that's a very good point. 00:21:14.500 --> 00:21:19.030 align:middle line:84% When I said before that you give P to the highest thing with 00:21:19.030 --> 00:21:21.880 align:middle line:84% a given label, I should have said something more 00:21:21.880 --> 00:21:27.400 align:middle line:84% sophisticated-- something like you give P to the highest thing 00:21:27.400 --> 00:21:31.090 align:middle line:84% with a given label that is produced or that is being 00:21:31.090 --> 00:21:33.550 align:middle line:90% projected from a given head-- 00:21:33.550 --> 00:21:35.570 align:middle line:90% from a given word. 00:21:35.570 --> 00:21:40.270 align:middle line:84% So this-- so "my pajamas" is an NP 00:21:40.270 --> 00:21:43.150 align:middle line:84% because it's the highest node that has an N that comes from 00:21:43.150 --> 00:21:45.100 align:middle line:90% "pajamas." 00:21:45.100 --> 00:21:50.860 align:middle line:84% And then there's a higher NP that's right under the VP right 00:21:50.860 --> 00:21:52.780 align:middle line:84% that you merge together with the verb, 00:21:52.780 --> 00:21:56.710 align:middle line:84% and that's an NP that's got its label from "elephant." 00:21:56.710 --> 00:21:58.692 align:middle line:84% So there are two NPs here-- one of them 00:21:58.692 --> 00:22:01.150 align:middle line:84% coming from "pajamas" and the other coming from "elephant." 00:22:01.150 --> 00:22:02.200 align:middle line:90% That's a very good point. 00:22:02.200 --> 00:22:03.742 align:middle line:84% What you're pointing out is that when 00:22:03.742 --> 00:22:06.310 align:middle line:84% I said the highest thing with the label N, 00:22:06.310 --> 00:22:07.570 align:middle line:90% I was being too fast. 00:22:07.570 --> 00:22:09.640 align:middle line:84% It's the highest thing with a given label N, 00:22:09.640 --> 00:22:11.712 align:middle line:84% and we have to distinguish N's from each other. 00:22:11.712 --> 00:22:12.670 align:middle line:90% Ooh, lots of questions. 00:22:12.670 --> 00:22:13.240 align:middle line:90% Yes? 00:22:13.240 --> 00:22:15.040 align:middle line:84% AUDIENCE: Actually, I think the thing 00:22:15.040 --> 00:22:18.920 align:middle line:84% about [INAUDIBLE] about the [INAUDIBLE] simple are not 00:22:18.920 --> 00:22:19.840 align:middle line:90% likely [INAUDIBLE]. 00:22:19.840 --> 00:22:22.198 align:middle line:84% Because I didn't get the normal reading. 00:22:22.198 --> 00:22:23.240 align:middle line:90% NORVIN RICHARDS: Oh, yes. 00:22:23.240 --> 00:22:23.740 align:middle line:90% [LAUGHS] 00:22:23.740 --> 00:22:29.240 align:middle line:84% AUDIENCE: [INAUDIBLE] imagining this guy was shooting himself 00:22:29.240 --> 00:22:31.088 align:middle line:84% while there was an elephant [INAUDIBLE].. 00:22:31.088 --> 00:22:33.380 align:middle line:84% NORVIN RICHARDS: Oh, the elephant is inside his pajamas 00:22:33.380 --> 00:22:34.547 align:middle line:90% and he is also wearing them. 00:22:34.547 --> 00:22:37.225 align:middle line:90% Ah, yes. 00:22:37.225 --> 00:22:39.600 align:middle line:84% I don't know what kind of tree we want to build for that. 00:22:39.600 --> 00:22:40.892 align:middle line:90% That's an interesting question. 00:22:40.892 --> 00:22:44.940 align:middle line:84% Actually, we're going to develop some tests in a second. 00:22:44.940 --> 00:22:48.110 align:middle line:84% So far, all I've done is assert this-- 00:22:48.110 --> 00:22:49.790 align:middle line:90% assert a couple of things. 00:22:49.790 --> 00:22:54.570 align:middle line:84% That string of words ought to have two different structures. 00:22:54.570 --> 00:22:56.030 align:middle line:84% So that prepositional phrase ought 00:22:56.030 --> 00:22:58.670 align:middle line:84% to be able to modify either a noun phrase or a verb phrase. 00:22:58.670 --> 00:23:01.970 align:middle line:84% And we have the sense that that string of words is ambiguous. 00:23:01.970 --> 00:23:05.000 align:middle line:84% It can mean a couple of different things. 00:23:05.000 --> 00:23:06.770 align:middle line:84% I haven't yet given you any reason 00:23:06.770 --> 00:23:09.350 align:middle line:84% to believe that these structures really 00:23:09.350 --> 00:23:12.410 align:middle line:84% are the structures that are associated with those meanings. 00:23:12.410 --> 00:23:15.320 align:middle line:84% I'm going to try to do that in a second to motivate that claim. 00:23:15.320 --> 00:23:17.790 align:middle line:84% And let's bear your reading in mind as I do that, 00:23:17.790 --> 00:23:20.120 align:middle line:84% and we'll see how it behaves with respect to the tests 00:23:20.120 --> 00:23:21.500 align:middle line:90% that I'm going to show you. 00:23:21.500 --> 00:23:23.420 align:middle line:84% This should all feel kind of familiar, right? 00:23:23.420 --> 00:23:26.270 align:middle line:84% So when we were doing "unlockable," we were saying, 00:23:26.270 --> 00:23:28.040 align:middle line:84% yeah, there's three morphemes-- there are 00:23:28.040 --> 00:23:29.340 align:middle line:90% two ways you can combine them. 00:23:29.340 --> 00:23:32.120 align:middle line:84% And if we think about what these morphemes do with the things 00:23:32.120 --> 00:23:35.180 align:middle line:84% that they combine with, we can understand 00:23:35.180 --> 00:23:37.310 align:middle line:84% why the word "unlockable" is ambiguous in the way 00:23:37.310 --> 00:23:38.330 align:middle line:90% that it is. 00:23:38.330 --> 00:23:40.560 align:middle line:84% We're doing something similar here. 00:23:40.560 --> 00:23:43.910 align:middle line:84% So our rules for how things combine 00:23:43.910 --> 00:23:45.800 align:middle line:84% allow us to create these strings of words 00:23:45.800 --> 00:23:47.270 align:middle line:90% in a couple of different ways. 00:23:47.270 --> 00:23:51.770 align:middle line:84% And it's ambiguous-- wonder if our freedom of building 00:23:51.770 --> 00:23:52.430 align:middle line:90% couple-- 00:23:52.430 --> 00:23:54.590 align:middle line:84% having a couple of different ways to build trees for this 00:23:54.590 --> 00:23:56.480 align:middle line:84% has anything to do with the fact that it seems 00:23:56.480 --> 00:23:57.710 align:middle line:90% to mean two different things. 00:23:57.710 --> 00:23:59.720 align:middle line:90% Joseph, did you have a question? 00:23:59.720 --> 00:24:00.560 align:middle line:90% AUDIENCE: Yeah. 00:24:00.560 --> 00:24:03.775 align:middle line:84% So looking at-- when you said that-- 00:24:03.775 --> 00:24:05.150 align:middle line:84% because on the left tree, there's 00:24:05.150 --> 00:24:07.095 align:middle line:90% two different noun phrases. 00:24:07.095 --> 00:24:11.520 align:middle line:84% And so following up on that, if you have-- 00:24:11.520 --> 00:24:14.010 align:middle line:84% would it also be appropriate to say-- 00:24:14.010 --> 00:24:18.300 align:middle line:84% so "pajamas" down all the way on the right 00:24:18.300 --> 00:24:19.620 align:middle line:90% is part of a noun phrase-- 00:24:19.620 --> 00:24:21.660 align:middle line:84% and that's part of a prepositional phrase, 00:24:21.660 --> 00:24:24.020 align:middle line:84% so that kind of chain of the noun phrase is terminated, 00:24:24.020 --> 00:24:26.220 align:middle line:84% so it's OK to start over with that? 00:24:26.220 --> 00:24:28.560 align:middle line:84% NORVIN RICHARDS: That's-- yes, that would work fine too. 00:24:28.560 --> 00:24:32.640 align:middle line:84% Your version of this raises the question 00:24:32.640 --> 00:24:37.470 align:middle line:84% of whether we'll ever find a noun phrase that has-- 00:24:37.470 --> 00:24:40.320 align:middle line:84% that is merged together with something else 00:24:40.320 --> 00:24:43.220 align:middle line:84% that also has the label N, or whether we'll ever 00:24:43.220 --> 00:24:44.970 align:middle line:84% find a verb phrase that is merged together 00:24:44.970 --> 00:24:47.430 align:middle line:84% with something else that also has the label V. I don't know 00:24:47.430 --> 00:24:49.230 align:middle line:84% whether we'll get around to seeing examples 00:24:49.230 --> 00:24:50.380 align:middle line:90% like that in this class. 00:24:50.380 --> 00:24:52.260 align:middle line:84% It's not clear that there are such examples, 00:24:52.260 --> 00:24:54.900 align:middle line:84% and if there aren't, then that's interesting. 00:24:54.900 --> 00:24:57.100 align:middle line:84% We might want to make something of that. 00:24:57.100 --> 00:25:00.077 align:middle line:84% So yes, your amendment to what I said 00:25:00.077 --> 00:25:02.160 align:middle line:84% and my amendment to what I said are both good ways 00:25:02.160 --> 00:25:06.300 align:middle line:84% of talking about the answer to Raquel's question from earlier. 00:25:06.300 --> 00:25:07.176 align:middle line:90% Yeah? 00:25:07.176 --> 00:25:08.650 align:middle line:84% AUDIENCE: So just to follow-up, I 00:25:08.650 --> 00:25:12.423 align:middle line:84% know someone asked about "I will," and so that, of course, 00:25:12.423 --> 00:25:14.840 align:middle line:90% has a modifier. 00:25:14.840 --> 00:25:19.010 align:middle line:84% But "I" in the sentence, do we just not [INAUDIBLE]?? 00:25:19.010 --> 00:25:20.010 align:middle line:90% NORVIN RICHARDS: Oh, no. 00:25:20.010 --> 00:25:20.730 align:middle line:90% See-- I'm sorry. 00:25:20.730 --> 00:25:22.147 align:middle line:84% I'm being very unfair to you guys. 00:25:22.147 --> 00:25:24.717 align:middle line:84% I keep only showing you trees for parts of sentences. 00:25:24.717 --> 00:25:26.550 align:middle line:84% So we've not-- we haven't gotten any further 00:25:26.550 --> 00:25:27.885 align:middle line:90% than the verb phrase. 00:25:27.885 --> 00:25:29.010 align:middle line:90% AUDIENCE: That [INAUDIBLE]. 00:25:29.010 --> 00:25:31.830 align:middle line:84% NORVIN RICHARDS: So I'm trying to get as much mileage out 00:25:31.830 --> 00:25:34.170 align:middle line:84% of the verb phrase as I can before we go on 00:25:34.170 --> 00:25:35.367 align:middle line:90% to larger things. 00:25:35.367 --> 00:25:37.950 align:middle line:84% But you're right, eventually, we will do sentences, I promise. 00:25:37.950 --> 00:25:40.800 align:middle line:90% 00:25:40.800 --> 00:25:42.960 align:middle line:90% Questions about this? 00:25:42.960 --> 00:25:45.390 align:middle line:90% They're all good. 00:25:45.390 --> 00:25:48.240 align:middle line:84% I was saying to the TAs the other day, you guys are-- 00:25:48.240 --> 00:25:50.680 align:middle line:84% it's nice being in a class full of talkative people. 00:25:50.680 --> 00:25:53.370 align:middle line:90% I appreciate it. 00:25:53.370 --> 00:25:56.967 align:middle line:84% So two trees, and a hypothesis that we might entertain 00:25:56.967 --> 00:25:59.550 align:middle line:84% is that the tree on the left is the one where the elephant was 00:25:59.550 --> 00:26:01.425 align:middle line:84% wearing the pajamas and the tree on the right 00:26:01.425 --> 00:26:05.220 align:middle line:84% is the one where the shooting involves 00:26:05.220 --> 00:26:07.206 align:middle line:90% me being in the pajamas-- 00:26:07.206 --> 00:26:09.330 align:middle line:90% sort of the normal reading-- 00:26:09.330 --> 00:26:10.830 align:middle line:84% normal for some of us, although it's 00:26:10.830 --> 00:26:12.788 align:middle line:84% been established that not all of us are normal. 00:26:12.788 --> 00:26:15.260 align:middle line:90% 00:26:15.260 --> 00:26:21.040 align:middle line:84% Now, remember back when I was trying to convince you 00:26:21.040 --> 00:26:23.500 align:middle line:84% that there were such things as constituents-- 00:26:23.500 --> 00:26:31.090 align:middle line:84% that syntax cared about whether a particular string of words 00:26:31.090 --> 00:26:36.130 align:middle line:84% were all descended from a single node or not. 00:26:36.130 --> 00:26:39.130 align:middle line:84% I was showing you tests for constituent structure, 00:26:39.130 --> 00:26:41.650 align:middle line:84% and in your recitation sessions, you 00:26:41.650 --> 00:26:43.900 align:middle line:84% may have played around with different tests 00:26:43.900 --> 00:26:45.460 align:middle line:90% for constituent structure. 00:26:45.460 --> 00:26:48.880 align:middle line:84% So one of the tests that we were fooling around with 00:26:48.880 --> 00:26:52.920 align:middle line:84% was what I called topicalization. 00:26:52.920 --> 00:26:56.170 align:middle line:84% It's possible to emphasize something 00:26:56.170 --> 00:26:59.750 align:middle line:84% by moving it to the beginning of the sentence. 00:26:59.750 --> 00:27:04.330 align:middle line:84% So you can say things like "the elephant I shot in my pajamas," 00:27:04.330 --> 00:27:09.160 align:middle line:84% or "the elephant in my pajamas I shot." 00:27:09.160 --> 00:27:13.825 align:middle line:84% Now, let's consider the tree on the left-- 00:27:13.825 --> 00:27:16.870 align:middle line:90% 00:27:16.870 --> 00:27:17.560 align:middle line:90% this tree. 00:27:17.560 --> 00:27:20.530 align:middle line:84% Is there a constituent in that tree-- 00:27:20.530 --> 00:27:27.120 align:middle line:84% an "elephant" that doesn't have anything else in it? 00:27:27.120 --> 00:27:30.240 align:middle line:84% Several of you are appropriately shaking your heads. 00:27:30.240 --> 00:27:37.927 align:middle line:84% There are nodes that are above the words "an elephant." 00:27:37.927 --> 00:27:39.510 align:middle line:84% There's a D that's above "in," there's 00:27:39.510 --> 00:27:40.770 align:middle line:90% an N that's above "elephant." 00:27:40.770 --> 00:27:44.730 align:middle line:84% And there's an NP up there which has 00:27:44.730 --> 00:27:48.760 align:middle line:84% "an elephant" as part of it, but it also has "in my pajamas." 00:27:48.760 --> 00:27:50.310 align:middle line:84% So in the tree on the left, there's 00:27:50.310 --> 00:27:53.080 align:middle line:90% no constituent "an elephant." 00:27:53.080 --> 00:27:54.580 align:middle line:84% How about in the tree on the right-- 00:27:54.580 --> 00:27:57.010 align:middle line:84% is there a constituent "an elephant"? 00:27:57.010 --> 00:27:59.390 align:middle line:84% Yeah, there's a noun phrase like that. 00:27:59.390 --> 00:28:01.190 align:middle line:84% So when we do the topicalization-- 00:28:01.190 --> 00:28:03.460 align:middle line:84% "the elephant I shot in my pajamas"-- 00:28:03.460 --> 00:28:06.130 align:middle line:84% we should only be able to do the tree on the right and not 00:28:06.130 --> 00:28:07.730 align:middle line:90% the tree on the left. 00:28:07.730 --> 00:28:09.670 align:middle line:84% That is, it should only have the meaning 00:28:09.670 --> 00:28:11.770 align:middle line:90% that the tree on the right has-- 00:28:11.770 --> 00:28:15.042 align:middle line:84% that I was wearing the pajamas and not the elephant. 00:28:15.042 --> 00:28:16.750 align:middle line:84% And several of you are raising your hands 00:28:16.750 --> 00:28:20.690 align:middle line:84% and I'll call on you in just a second, but before I do that, 00:28:20.690 --> 00:28:23.690 align:middle line:84% do people have the feeling that that's true? 00:28:23.690 --> 00:28:26.690 align:middle line:84% That if I say "The elephant I shot in my pajamas-- 00:28:26.690 --> 00:28:30.890 align:middle line:84% the aardvark I shot in my tuxedo," 00:28:30.890 --> 00:28:32.330 align:middle line:90% that I'm wearing these things. 00:28:32.330 --> 00:28:34.850 align:middle line:90% It's not my victims. 00:28:34.850 --> 00:28:36.470 align:middle line:90% And I think that's right. 00:28:36.470 --> 00:28:39.140 align:middle line:84% So the ambiguity that we have if we haven't 00:28:39.140 --> 00:28:40.880 align:middle line:90% done any topicalization-- 00:28:40.880 --> 00:28:43.650 align:middle line:84% that ambiguity was there because we had both of these trees. 00:28:43.650 --> 00:28:46.550 align:middle line:84% And if we run a constituency test, 00:28:46.550 --> 00:28:49.788 align:middle line:84% that collapses the ambiguity, because it makes it 00:28:49.788 --> 00:28:51.830 align:middle line:84% so that we could only have the tree on the right, 00:28:51.830 --> 00:28:52.770 align:middle line:90% not the tree on the left. 00:28:52.770 --> 00:28:54.770 align:middle line:84% And so we can only have the meaning on the right 00:28:54.770 --> 00:28:56.510 align:middle line:90% and not the meaning on the left. 00:28:56.510 --> 00:28:57.560 align:middle line:90% Yes? 00:28:57.560 --> 00:28:59.878 align:middle line:84% Sorry, I'll get to you in just a second. 00:28:59.878 --> 00:29:01.670 align:middle line:84% AUDIENCE: So yeah, on the tree on the left, 00:29:01.670 --> 00:29:04.610 align:middle line:84% would it be more correct to combine "an elephant" first 00:29:04.610 --> 00:29:07.810 align:middle line:84% before adding it to the big upper tree? 00:29:07.810 --> 00:29:10.480 align:middle line:84% NORVIN RICHARDS: Oh, you mean combine "an elephant" first 00:29:10.480 --> 00:29:13.850 align:middle line:84% and then put "in my pajamas" together with that? 00:29:13.850 --> 00:29:16.420 align:middle line:90% AUDIENCE: Yeah. 00:29:16.420 --> 00:29:18.850 align:middle line:84% NORVIN RICHARDS: So we could do that. 00:29:18.850 --> 00:29:24.850 align:middle line:84% Notice that if we did that, there would again 00:29:24.850 --> 00:29:27.800 align:middle line:90% be a constituent "an elephant." 00:29:27.800 --> 00:29:30.280 align:middle line:84% So we would lose the result that we've 00:29:30.280 --> 00:29:33.280 align:middle line:84% got here, which is that if you front "the elephant"-- 00:29:33.280 --> 00:29:35.867 align:middle line:84% I'm sorry, I keep switching back and forth between the and an. 00:29:35.867 --> 00:29:38.200 align:middle line:84% I'll just ask you to believe me that they're the same as 00:29:38.200 --> 00:29:41.170 align:middle line:90% far as this is concerned. 00:29:41.170 --> 00:29:42.788 align:middle line:84% But if we topicalize "the elephant"-- 00:29:42.788 --> 00:29:44.830 align:middle line:84% if we move that to the beginning of the sentence, 00:29:44.830 --> 00:29:46.570 align:middle line:84% then we lose the tree on the left. 00:29:46.570 --> 00:29:49.000 align:middle line:84% So you're-- let me make sure I'm understanding what 00:29:49.000 --> 00:29:49.640 align:middle line:90% you're saying. 00:29:49.640 --> 00:29:52.140 align:middle line:84% You're saying, wait-- what about a tree that would look like 00:29:52.140 --> 00:29:53.290 align:middle line:90% this-- 00:29:53.290 --> 00:30:04.630 align:middle line:84% shot, and then we've got a noun phrase "an elephant," 00:30:04.630 --> 00:30:07.120 align:middle line:84% and then we've got a prepositional phrase here 00:30:07.120 --> 00:30:11.350 align:middle line:84% "in the (noun phrase)," and then determiner "my," 00:30:11.350 --> 00:30:14.920 align:middle line:90% and then noun "pajamas." 00:30:14.920 --> 00:30:17.980 align:middle line:90% You're imagining this tree. 00:30:17.980 --> 00:30:21.760 align:middle line:84% And yes, notice, though, that if we had that tree, 00:30:21.760 --> 00:30:25.660 align:middle line:84% there would be a constituent here-- "an elephant." 00:30:25.660 --> 00:30:27.460 align:middle line:84% So now, there are two ways we can go. 00:30:27.460 --> 00:30:30.340 align:middle line:84% One would be to say aha, we're learning 00:30:30.340 --> 00:30:34.090 align:middle line:84% that although prepositional phrase can merge 00:30:34.090 --> 00:30:38.200 align:middle line:90% with a projection of a noun-- 00:30:38.200 --> 00:30:39.700 align:middle line:90% something with the label N-- 00:30:39.700 --> 00:30:42.520 align:middle line:84% it can't-- it has to merge with N, 00:30:42.520 --> 00:30:46.330 align:middle line:84% or maybe it has to merge before you merge a D-- 00:30:46.330 --> 00:30:47.830 align:middle line:84% that there are rules about the order 00:30:47.830 --> 00:30:49.270 align:middle line:90% in which you merge things. 00:30:49.270 --> 00:30:51.880 align:middle line:84% We're going to get a chance to talk about things like that 00:30:51.880 --> 00:30:52.540 align:middle line:90% soon-- 00:30:52.540 --> 00:30:56.920 align:middle line:84% other places where you are so far-- let's see. 00:30:56.920 --> 00:31:01.390 align:middle line:84% We're at the stage of syntax where life is easy and free. 00:31:01.390 --> 00:31:04.360 align:middle line:84% There's ambiguity-- you can create trees however you want. 00:31:04.360 --> 00:31:08.770 align:middle line:84% You've now come up with a tree that we want to exclude somehow 00:31:08.770 --> 00:31:12.370 align:middle line:84% in order to avoid the result that first sentence would have 00:31:12.370 --> 00:31:14.950 align:middle line:90% this tree as a possible tree. 00:31:14.950 --> 00:31:16.160 align:middle line:90% We want to avoid that. 00:31:16.160 --> 00:31:17.440 align:middle line:84% So there are two kinds of things we could do. 00:31:17.440 --> 00:31:19.810 align:middle line:84% One would be to say, no, you may not draw this tree, 00:31:19.810 --> 00:31:22.270 align:middle line:84% and maybe that would be about the order in which you 00:31:22.270 --> 00:31:24.640 align:middle line:84% can merge D and prepositional phrases. 00:31:24.640 --> 00:31:26.470 align:middle line:84% That would be one thing we could do. 00:31:26.470 --> 00:31:27.887 align:middle line:84% Another thing we could do would be 00:31:27.887 --> 00:31:32.170 align:middle line:84% to say, yeah, in order to be topicalized, 00:31:32.170 --> 00:31:33.970 align:middle line:90% you have to be a constituent. 00:31:33.970 --> 00:31:36.550 align:middle line:84% But actually, there are some constituents that cannot 00:31:36.550 --> 00:31:39.730 align:middle line:84% topicalize-- that is, it's not a bi-conditional. 00:31:39.730 --> 00:31:41.830 align:middle line:84% That would be the other move to make. 00:31:41.830 --> 00:31:45.910 align:middle line:84% So this has been a very long and elaborate version of "yes." 00:31:45.910 --> 00:31:48.970 align:middle line:84% And so you're right-- this is an imaginable tree, 00:31:48.970 --> 00:31:50.618 align:middle line:90% and so we must do something. 00:31:50.618 --> 00:31:52.660 align:middle line:84% So I showed you a couple of things that we could. 00:31:52.660 --> 00:31:54.700 align:middle line:90% Does that make sense at all? 00:31:54.700 --> 00:31:55.240 align:middle line:90% OK? 00:31:55.240 --> 00:31:55.930 align:middle line:90% Good. 00:31:55.930 --> 00:31:57.040 align:middle line:90% Katrina? 00:31:57.040 --> 00:31:59.140 align:middle line:84% AUDIENCE: I was also going to bring up 00:31:59.140 --> 00:32:02.800 align:middle line:84% that you could combine "an elephant" before merging 00:32:02.800 --> 00:32:04.973 align:middle line:84% to the prepositional phrase, and so-- 00:32:04.973 --> 00:32:06.140 align:middle line:90% NORVIN RICHARDS: Good point. 00:32:06.140 --> 00:32:07.570 align:middle line:90% And so I've answered. 00:32:07.570 --> 00:32:08.440 align:middle line:90% Good-- excellent. 00:32:08.440 --> 00:32:09.040 align:middle line:90% Good question. 00:32:09.040 --> 00:32:09.820 align:middle line:90% Yes, Joseph? 00:32:09.820 --> 00:32:12.470 align:middle line:84% AUDIENCE: This might be getting ahead of this-- 00:32:12.470 --> 00:32:14.750 align:middle line:84% NORVIN RICHARDS: Where we want to be? 00:32:14.750 --> 00:32:18.015 align:middle line:84% AUDIENCE: --but obviously, I [INAUDIBLE].. 00:32:18.015 --> 00:32:21.715 align:middle line:84% So if you break that constituent off "an elephant" 00:32:21.715 --> 00:32:27.490 align:middle line:84% in the second tree corresponding to the first [INAUDIBLE].. 00:32:27.490 --> 00:32:31.490 align:middle line:90% Breaking off "an elephant"-- 00:32:31.490 --> 00:32:34.850 align:middle line:84% I don't know what the actual topology of these trees are 00:32:34.850 --> 00:32:37.240 align:middle line:84% and what the rules are, but aren't you 00:32:37.240 --> 00:32:39.782 align:middle line:90% crossing that over the I branch? 00:32:39.782 --> 00:32:41.990 align:middle line:84% NORVIN RICHARDS: Why don't you remember that question 00:32:41.990 --> 00:32:45.560 align:middle line:84% and ask it again once we have complete sentences? 00:32:45.560 --> 00:32:49.550 align:middle line:84% I mean, so far, the only condition on topicalization-- 00:32:49.550 --> 00:32:51.073 align:middle line:84% this operation that puts something 00:32:51.073 --> 00:32:52.490 align:middle line:84% at the beginning of the sentence-- 00:32:52.490 --> 00:32:55.280 align:middle line:84% that I've offered you is that the thing that you topicalize 00:32:55.280 --> 00:32:57.500 align:middle line:90% has to be a constituent. 00:32:57.500 --> 00:32:59.750 align:middle line:84% I haven't said anything else about where-- 00:32:59.750 --> 00:33:01.220 align:middle line:84% like, I haven't offered to draw you 00:33:01.220 --> 00:33:04.010 align:middle line:84% a tree for "the elephant I shot in my pajamas," for example. 00:33:04.010 --> 00:33:07.410 align:middle line:84% Eventually, I will, but I'm not going to do that yet. 00:33:07.410 --> 00:33:10.278 align:middle line:84% I may not even do it today, partly because eventually, 00:33:10.278 --> 00:33:12.320 align:middle line:84% we have to get "I shot an elephant in my pajamas" 00:33:12.320 --> 00:33:13.115 align:middle line:90% just for an example. 00:33:13.115 --> 00:33:13.615 align:middle line:90% [SNEEZING] 00:33:13.615 --> 00:33:14.480 align:middle line:90% Bless you. 00:33:14.480 --> 00:33:17.420 align:middle line:90% So we'll get that first. 00:33:17.420 --> 00:33:19.177 align:middle line:90% Yes? 00:33:19.177 --> 00:33:21.010 align:middle line:84% AUDIENCE: I know this is kind of [INAUDIBLE] 00:33:21.010 --> 00:33:25.280 align:middle line:84% but with the example of "an elephant" [INAUDIBLE] 00:33:25.280 --> 00:33:29.150 align:middle line:84% is that its own constituent so that we could technically 00:33:29.150 --> 00:33:34.367 align:middle line:84% be like, "the elephant, comma, in my pajamas, comma"? 00:33:34.367 --> 00:33:35.450 align:middle line:90% NORVIN RICHARDS: Oh, dear. 00:33:35.450 --> 00:33:37.610 align:middle line:84% AUDIENCE: Like, separated it if was [INAUDIBLE].. 00:33:37.610 --> 00:33:38.777 align:middle line:90% NORVIN RICHARDS: If it was-- 00:33:38.777 --> 00:33:40.820 align:middle line:84% you mean if we were able to draw trees like this? 00:33:40.820 --> 00:33:41.445 align:middle line:90% AUDIENCE: Yeah. 00:33:41.445 --> 00:33:43.628 align:middle line:90% 00:33:43.628 --> 00:33:45.170 align:middle line:84% NORVIN RICHARDS: I shot an elephant-- 00:33:45.170 --> 00:33:46.432 align:middle line:90% AUDIENCE: Oh, wait. 00:33:46.432 --> 00:33:48.052 align:middle line:90% Did we explain why couldn't-- 00:33:48.052 --> 00:33:50.010 align:middle line:84% NORVIN RICHARDS: Why we had better not do this? 00:33:50.010 --> 00:33:53.872 align:middle line:90% So we had two people point out-- 00:33:53.872 --> 00:33:55.580 align:middle line:84% one of them actually got to point it out, 00:33:55.580 --> 00:33:57.320 align:middle line:84% and then Kateryna pointed out that she 00:33:57.320 --> 00:33:58.580 align:middle line:90% was going to point it out-- 00:33:58.580 --> 00:34:01.250 align:middle line:84% that this is a tree that you could imagine. 00:34:01.250 --> 00:34:03.810 align:middle line:84% You could imagine being able to build a tree like this. 00:34:03.810 --> 00:34:06.290 align:middle line:90% And my reaction was, oh, dear-- 00:34:06.290 --> 00:34:08.690 align:middle line:84% we must do something to stop that. 00:34:08.690 --> 00:34:11.748 align:middle line:84% So here-- I'll mark this tree with a frowny face 00:34:11.748 --> 00:34:12.290 align:middle line:90% or something. 00:34:12.290 --> 00:34:15.650 align:middle line:84% We've got to exclude this tree somehow-- 00:34:15.650 --> 00:34:16.969 align:middle line:90% maybe. 00:34:16.969 --> 00:34:20.510 align:middle line:84% We've got we've got to avoid the following problem. 00:34:20.510 --> 00:34:22.699 align:middle line:84% This sure looks like a tree for a meaning 00:34:22.699 --> 00:34:24.650 align:middle line:84% where the elephant is wearing the pajamas. 00:34:24.650 --> 00:34:26.210 align:middle line:84% So a tree that has the same meaning 00:34:26.210 --> 00:34:27.949 align:middle line:90% as the tree on the left. 00:34:27.949 --> 00:34:29.540 align:middle line:84% But it's also a tree in which there's 00:34:29.540 --> 00:34:31.429 align:middle line:90% a constituent "an elephant." 00:34:31.429 --> 00:34:33.500 align:middle line:84% And the fact that I wanted to get across 00:34:33.500 --> 00:34:35.300 align:middle line:84% to you with this slide was that if you 00:34:35.300 --> 00:34:37.120 align:middle line:90% use topicalization to make it-- 00:34:37.120 --> 00:34:40.010 align:middle line:84% to find out-- I was only going to choose between these two 00:34:40.010 --> 00:34:40.550 align:middle line:90% trees-- 00:34:40.550 --> 00:34:43.920 align:middle line:84% to find out whether an elephant is a constituent or not, 00:34:43.920 --> 00:34:46.050 align:middle line:84% then you have to have the tree on the right 00:34:46.050 --> 00:34:47.300 align:middle line:90% where I'm wearing the pajamas. 00:34:47.300 --> 00:34:49.333 align:middle line:84% The elephant can't be wearing the pajamas. 00:34:49.333 --> 00:34:50.750 align:middle line:84% Two people immediately said, wait, 00:34:50.750 --> 00:34:52.750 align:middle line:84% what about this-- imagine the frowning face tree 00:34:52.750 --> 00:34:54.770 align:middle line:90% that I don't have on the slide. 00:34:54.770 --> 00:34:56.690 align:middle line:84% And I said there are two things we could do. 00:34:56.690 --> 00:35:01.160 align:middle line:84% One would be to say, no, you may not do this maybe because there 00:35:01.160 --> 00:35:04.580 align:middle line:84% are rules-- which maybe someday, we'll get to-- that constrain 00:35:04.580 --> 00:35:08.690 align:middle line:84% the order in which you can merge a prepositional phrase and a D 00:35:08.690 --> 00:35:10.010 align:middle line:90% with a noun-- 00:35:10.010 --> 00:35:12.920 align:middle line:84% rules that will guarantee that if you're 00:35:12.920 --> 00:35:15.230 align:middle line:84% going to put a prepositional phrase and also a D 00:35:15.230 --> 00:35:17.180 align:middle line:84% in projections of a noun, you better 00:35:17.180 --> 00:35:19.250 align:middle line:84% do the prepositional phrase first and then the D. 00:35:19.250 --> 00:35:21.650 align:middle line:84% And then we get to ask, well, why? 00:35:21.650 --> 00:35:22.980 align:middle line:90% Where did that rule come from? 00:35:22.980 --> 00:35:25.105 align:middle line:84% And the answer is, well, this is day two of syntax. 00:35:25.105 --> 00:35:26.270 align:middle line:90% But we'll get there. 00:35:26.270 --> 00:35:27.290 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:35:27.290 --> 00:35:29.040 align:middle line:84% NORVIN RICHARDS: Did that-- no, no, no, that's fine. 00:35:29.040 --> 00:35:30.470 align:middle line:84% The other thing we could do-- and here, 00:35:30.470 --> 00:35:32.070 align:middle line:84% let me say something I said before, 00:35:32.070 --> 00:35:34.195 align:middle line:84% but I'll say it a little more coherently possibly-- 00:35:34.195 --> 00:35:34.940 align:middle line:90% we'll see-- 00:35:34.940 --> 00:35:39.260 align:middle line:84% would be to say in order to topicalize-- in order 00:35:39.260 --> 00:35:41.330 align:middle line:84% to be fronted-- you have to be a constituent. 00:35:41.330 --> 00:35:43.250 align:middle line:84% But it's not enough to be a constituent. 00:35:43.250 --> 00:35:45.920 align:middle line:84% Notice, the thing that you would be fronting 00:35:45.920 --> 00:35:47.540 align:middle line:90% in the tree on the right. 00:35:47.540 --> 00:35:51.437 align:middle line:84% It's not just a constituent, it's a noun phrase. 00:35:51.437 --> 00:35:54.020 align:middle line:84% And what you would be fronting over here is not a noun phrase, 00:35:54.020 --> 00:35:55.970 align:middle line:90% it's an N bar. 00:35:55.970 --> 00:35:58.670 align:middle line:84% So you could say, ah, we're learning something new 00:35:58.670 --> 00:36:00.635 align:middle line:90% about topicalization. 00:36:00.635 --> 00:36:03.453 align:middle line:84% Topicalization-- yes, you must be a constituent, 00:36:03.453 --> 00:36:05.120 align:middle line:84% but it's not enough to be a constituent. 00:36:05.120 --> 00:36:06.620 align:middle line:90% You must actually be a phrase. 00:36:06.620 --> 00:36:08.540 align:middle line:90% You cannot just be a bar. 00:36:08.540 --> 00:36:12.020 align:middle line:84% And then that would be a new and exciting thing to learn. 00:36:12.020 --> 00:36:13.940 align:middle line:84% We have to find out which of these ways 00:36:13.940 --> 00:36:16.068 align:middle line:84% of dealing with this tree is the better way. 00:36:16.068 --> 00:36:17.360 align:middle line:90% You had a question a while ago. 00:36:17.360 --> 00:36:17.860 align:middle line:90% I'm sorry. 00:36:17.860 --> 00:36:21.590 align:middle line:84% AUDIENCE: Yes, so I might be jumping the gun here, 00:36:21.590 --> 00:36:27.240 align:middle line:84% but it almost seems to me like in the first tree-- 00:36:27.240 --> 00:36:28.100 align:middle line:90% no, that's a lie. 00:36:28.100 --> 00:36:28.600 align:middle line:90% Never mind. 00:36:28.600 --> 00:36:29.517 align:middle line:90% NORVIN RICHARDS: Good. 00:36:29.517 --> 00:36:30.150 align:middle line:90% [LAUGHTER] 00:36:30.150 --> 00:36:31.265 align:middle line:90% Yes? 00:36:31.265 --> 00:36:33.390 align:middle line:84% AUDIENCE: We're doing all of this based on the fact 00:36:33.390 --> 00:36:36.630 align:middle line:84% that we don't want the kind of meaning where the elephant is 00:36:36.630 --> 00:36:39.450 align:middle line:84% wearing your pajamas, but what if you want 00:36:39.450 --> 00:36:42.002 align:middle line:90% to say a sentence like that? 00:36:42.002 --> 00:36:42.960 align:middle line:90% NORVIN RICHARDS: Well-- 00:36:42.960 --> 00:36:45.060 align:middle line:84% AUDIENCE: We're trying to exclude 00:36:45.060 --> 00:36:46.792 align:middle line:84% that meaning from the sentence, but what 00:36:46.792 --> 00:36:48.000 align:middle line:90% if someone wants to say that. 00:36:48.000 --> 00:36:48.890 align:middle line:90% NORVIN RICHARDS: Well, so-- 00:36:48.890 --> 00:36:50.890 align:middle line:84% AUDIENCE: "I exterminated the bugs in my walls." 00:36:50.890 --> 00:36:52.063 align:middle line:90% NORVIN RICHARDS: So wait. 00:36:52.063 --> 00:36:53.480 align:middle line:84% Leave the bugs alone for a second. 00:36:53.480 --> 00:36:55.080 align:middle line:90% [LAUGHTER] 00:36:55.080 --> 00:36:58.220 align:middle line:84% We have-- there's enough death in this class already. 00:36:58.220 --> 00:36:59.000 align:middle line:90% Whoa, go back. 00:36:59.000 --> 00:37:02.480 align:middle line:90% 00:37:02.480 --> 00:37:04.610 align:middle line:90% If somebody wants the meaning-- 00:37:04.610 --> 00:37:05.318 align:middle line:90% so let's back up. 00:37:05.318 --> 00:37:07.568 align:middle line:84% Here's what I was going to do with this pair of trees. 00:37:07.568 --> 00:37:09.950 align:middle line:84% I was going to say, hey, look, "I shot the elephant 00:37:09.950 --> 00:37:11.750 align:middle line:84% in my pajamas" is ambiguous in the way 00:37:11.750 --> 00:37:13.370 align:middle line:84% that Groucho Marx took advantage of, 00:37:13.370 --> 00:37:15.120 align:middle line:90% and the ambiguity goes away. 00:37:15.120 --> 00:37:17.120 align:middle line:84% And so far, we've only done the first sentence-- 00:37:17.120 --> 00:37:18.770 align:middle line:84% "The elephant I shot in my pajamas." 00:37:18.770 --> 00:37:20.390 align:middle line:84% That has to be the tree on the right. 00:37:20.390 --> 00:37:21.973 align:middle line:84% The tree on the right is the only tree 00:37:21.973 --> 00:37:23.312 align:middle line:90% in which that's a constituent. 00:37:23.312 --> 00:37:25.520 align:middle line:84% Immediately, people began doing the frowny face tree. 00:37:25.520 --> 00:37:27.228 align:middle line:84% Forget the frowny face tree for a second. 00:37:27.228 --> 00:37:29.165 align:middle line:90% Those are the only two trees. 00:37:29.165 --> 00:37:30.950 align:middle line:84% It has to be the tree on the right. 00:37:30.950 --> 00:37:36.600 align:middle line:84% And "The elephant in my pajamas I shot"-- 00:37:36.600 --> 00:37:40.998 align:middle line:84% the ambiguity also collapses, but in the other direction. 00:37:40.998 --> 00:37:42.040 align:middle line:90% People have that feeling? 00:37:42.040 --> 00:37:43.920 align:middle line:84% So if I say, "The elephant in my pajamas 00:37:43.920 --> 00:37:49.830 align:middle line:84% I shot-- the elephant in my tuxedo I offered some bananas." 00:37:49.830 --> 00:37:52.950 align:middle line:90% 00:37:52.950 --> 00:37:54.450 align:middle line:84% That doesn't mean I shot an elephant 00:37:54.450 --> 00:37:55.742 align:middle line:90% while I was wearing my pajamas. 00:37:55.742 --> 00:37:58.080 align:middle line:84% That second one-- "The elephant in my pajamas I shot"-- 00:37:58.080 --> 00:38:00.370 align:middle line:84% only means that the elephant was wearing the pajamas. 00:38:00.370 --> 00:38:00.930 align:middle line:90% Why? 00:38:00.930 --> 00:38:03.120 align:middle line:84% Well, if we look at the trees we can see why. 00:38:03.120 --> 00:38:05.820 align:middle line:84% There's a noun phrase-- "an elephant in my pajamas"-- 00:38:05.820 --> 00:38:07.620 align:middle line:84% in the tree on the left, and there's 00:38:07.620 --> 00:38:09.720 align:middle line:84% no constituent "an elephant in my pajamas" 00:38:09.720 --> 00:38:10.810 align:middle line:90% in the tree on the right. 00:38:10.810 --> 00:38:13.260 align:middle line:90% So these constituency tests-- 00:38:13.260 --> 00:38:17.340 align:middle line:84% topicalization-- forces us to one or the other of those two 00:38:17.340 --> 00:38:18.060 align:middle line:90% trees. 00:38:18.060 --> 00:38:20.955 align:middle line:84% And depending on which tree you're forced into, 00:38:20.955 --> 00:38:24.285 align:middle line:84% you only have the reading that's associated with that tree. 00:38:24.285 --> 00:38:27.280 align:middle line:84% And so the ambiguity collapses in the way that it should. 00:38:27.280 --> 00:38:29.710 align:middle line:84% So once we nail down which tree we're looking at, 00:38:29.710 --> 00:38:31.740 align:middle line:84% we also nail down which meaning we have. 00:38:31.740 --> 00:38:35.930 align:middle line:90% 00:38:35.930 --> 00:38:37.580 align:middle line:84% Kateryna wanted to know what if I'm 00:38:37.580 --> 00:38:39.290 align:middle line:84% interested in the tree on the left? 00:38:39.290 --> 00:38:42.990 align:middle line:84% The answer is, well, either don't topicalize or do 00:38:42.990 --> 00:38:44.990 align:middle line:84% the topicalization-- the second topicalization-- 00:38:44.990 --> 00:38:46.820 align:middle line:84% The elephant in my pajamas I shot." 00:38:46.820 --> 00:38:52.160 align:middle line:84% Those are both consistent with the tree on the left, 00:38:52.160 --> 00:38:55.520 align:middle line:84% but you can't do the first topicalization. 00:38:55.520 --> 00:38:58.940 align:middle line:84% Now, have I already warned you about this? 00:38:58.940 --> 00:39:01.515 align:middle line:84% There's a danger if you go further in linguistics 00:39:01.515 --> 00:39:03.890 align:middle line:84% that you will lose the ability to distinguish grammatical 00:39:03.890 --> 00:39:05.270 align:middle line:90% from ungrammatical sentences. 00:39:05.270 --> 00:39:07.520 align:middle line:84% So your feelings about sentences will all 00:39:07.520 --> 00:39:09.050 align:middle line:90% become this kind of gray blur. 00:39:09.050 --> 00:39:11.210 align:middle line:84% It's sometimes called syntacticians' disease. 00:39:11.210 --> 00:39:15.470 align:middle line:84% And I have been a syntactician for longer than you have been 00:39:15.470 --> 00:39:17.690 align:middle line:90% alive, which is depressing. 00:39:17.690 --> 00:39:20.000 align:middle line:84% And so I no longer have any judgments 00:39:20.000 --> 00:39:21.560 align:middle line:90% at all about sentences. 00:39:21.560 --> 00:39:23.060 align:middle line:84% But I think this is how it works. 00:39:23.060 --> 00:39:24.950 align:middle line:84% Does anybody want to object when I 00:39:24.950 --> 00:39:27.140 align:middle line:84% say these topicalizations only have 00:39:27.140 --> 00:39:29.510 align:middle line:90% one meaning and not the other? 00:39:29.510 --> 00:39:30.534 align:middle line:90% Yeah? 00:39:30.534 --> 00:39:32.350 align:middle line:84% AUDIENCE: So how would you say that, 00:39:32.350 --> 00:39:37.160 align:middle line:84% while this topicalization [INAUDIBLE],, 00:39:37.160 --> 00:39:41.460 align:middle line:84% and while [INAUDIBLE],, it doesn't-- 00:39:41.460 --> 00:39:45.676 align:middle line:84% the first sentence still points to the second tree 00:39:45.676 --> 00:39:48.690 align:middle line:90% on the top slide. 00:39:48.690 --> 00:39:50.780 align:middle line:84% NORVIN RICHARDS: To the tree on the left? 00:39:50.780 --> 00:39:53.840 align:middle line:84% AUDIENCE: Right, because while they both have 00:39:53.840 --> 00:39:57.424 align:middle line:84% the constituent "an elephant," the difference 00:39:57.424 --> 00:40:00.310 align:middle line:90% is that "shot" has a verb-- 00:40:00.310 --> 00:40:04.700 align:middle line:84% has a V-bar that's combined with "in my pajamas" 00:40:04.700 --> 00:40:10.436 align:middle line:84% while "shot" is never directly combined with [INAUDIBLE].. 00:40:10.436 --> 00:40:12.623 align:middle line:84% So I only point out that there might 00:40:12.623 --> 00:40:17.726 align:middle line:84% be a rule that only when they're directly combined together, 00:40:17.726 --> 00:40:22.100 align:middle line:90% they have [INAUDIBLE]. 00:40:22.100 --> 00:40:24.890 align:middle line:84% NORVIN RICHARDS: So I was with you right up until-- 00:40:24.890 --> 00:40:29.240 align:middle line:84% I thought I was with you right up until the end there. 00:40:29.240 --> 00:40:33.920 align:middle line:84% The sentence, "I shot an elephant in my pajamas," 00:40:33.920 --> 00:40:38.120 align:middle line:84% the verb phrase can have either of these structures. 00:40:38.120 --> 00:40:39.125 align:middle line:90% Is that-- no? 00:40:39.125 --> 00:40:41.034 align:middle line:84% AUDIENCE: Because I think the first sentence, 00:40:41.034 --> 00:40:43.500 align:middle line:90% you said I shot in my pajamas. 00:40:43.500 --> 00:40:46.868 align:middle line:84% So it was shot [INAUDIBLE] pajamas [INAUDIBLE].. 00:40:46.868 --> 00:40:49.048 align:middle line:90% NORVIN RICHARDS: Oh, I see. 00:40:49.048 --> 00:40:51.340 align:middle line:84% And so you can only do that with the tree on the right, 00:40:51.340 --> 00:40:53.200 align:middle line:84% because "in my pajamas" got merged together 00:40:53.200 --> 00:40:56.830 align:middle line:84% with a projection of the verb "shot," whereas 00:40:56.830 --> 00:40:58.840 align:middle line:90% in the second one-- 00:40:58.840 --> 00:41:00.595 align:middle line:84% "The elephant in my pajamas I shot"-- 00:41:00.595 --> 00:41:03.370 align:middle line:90% 00:41:03.370 --> 00:41:08.294 align:middle line:84% what's wrong with having either of these trees? 00:41:08.294 --> 00:41:10.670 align:middle line:84% AUDIENCE: So "The elephant in my pajamas I shot" 00:41:10.670 --> 00:41:17.180 align:middle line:84% couldn't mean [INAUDIBLE] because an elephant is never 00:41:17.180 --> 00:41:19.770 align:middle line:90% [INAUDIBLE]. 00:41:19.770 --> 00:41:21.710 align:middle line:90% NORVIN RICHARDS: Ah, I see. 00:41:21.710 --> 00:41:23.360 align:middle line:90% Yes, I think-- 00:41:23.360 --> 00:41:24.920 align:middle line:84% I could be wrong, but I think what 00:41:24.920 --> 00:41:29.420 align:middle line:84% you are saying is another way of what I am saying. 00:41:29.420 --> 00:41:31.960 align:middle line:84% I think we're saying the same thing in different ways. 00:41:31.960 --> 00:41:36.110 align:middle line:84% So I am talking about these trees 00:41:36.110 --> 00:41:40.460 align:middle line:84% as though they are objects that you can grab nodes in them 00:41:40.460 --> 00:41:41.990 align:middle line:90% and move them around. 00:41:41.990 --> 00:41:43.520 align:middle line:84% You are talking about these trees 00:41:43.520 --> 00:41:46.700 align:middle line:84% as if they are sequences of events. 00:41:46.700 --> 00:41:49.430 align:middle line:84% They're-- which is the way, I encourage you to talk about 00:41:49.430 --> 00:41:50.270 align:middle line:90% trees earlier on. 00:41:50.270 --> 00:41:52.910 align:middle line:84% They're records of the order in which you 00:41:52.910 --> 00:41:54.890 align:middle line:90% put two things together. 00:41:54.890 --> 00:41:59.120 align:middle line:84% And you're saying, yeah, if you put "in my pajamas" 00:41:59.120 --> 00:42:03.030 align:middle line:84% together with "shot," then you can do the first sentence. 00:42:03.030 --> 00:42:04.700 align:middle line:84% And if you put "in my pajamas" together 00:42:04.700 --> 00:42:08.480 align:middle line:84% with "the elephant," then you can do the second sentence. 00:42:08.480 --> 00:42:11.690 align:middle line:84% That is, which kinds of things you can topicalize 00:42:11.690 --> 00:42:15.800 align:middle line:84% cares about which order you put things together in. 00:42:15.800 --> 00:42:18.810 align:middle line:84% And I think we are saying similar things. 00:42:18.810 --> 00:42:20.810 align:middle line:84% I'm just saying it representationally, 00:42:20.810 --> 00:42:21.350 align:middle line:90% if you want. 00:42:21.350 --> 00:42:23.420 align:middle line:84% I'm inviting you to look at these trees 00:42:23.420 --> 00:42:25.820 align:middle line:84% and think about them as objects that you can move around, 00:42:25.820 --> 00:42:28.670 align:middle line:84% and you're thinking in terms of the order in which you 00:42:28.670 --> 00:42:30.210 align:middle line:90% to put things together. 00:42:30.210 --> 00:42:33.050 align:middle line:84% So I think you're right, but you're right in the same way 00:42:33.050 --> 00:42:34.910 align:middle line:90% that I am right. 00:42:34.910 --> 00:42:37.780 align:middle line:84% This is the kind conclusion of this kind of conversation 00:42:37.780 --> 00:42:40.280 align:middle line:84% that I always like-- it's the one where everyone involved is 00:42:40.280 --> 00:42:42.300 align:middle line:90% right. 00:42:42.300 --> 00:42:46.140 align:middle line:84% Other questions about this so far? 00:42:46.140 --> 00:42:46.820 align:middle line:90% OK. 00:42:46.820 --> 00:42:49.310 align:middle line:84% So elephants-- is this enough elephant violence? 00:42:49.310 --> 00:42:50.150 align:middle line:90% Possibly. 00:42:50.150 --> 00:42:51.027 align:middle line:90% Let's see. 00:42:51.027 --> 00:42:52.610 align:middle line:84% Oh, no, there's so much more violence. 00:42:52.610 --> 00:42:53.810 align:middle line:90% I'm sorry. 00:42:53.810 --> 00:42:56.720 align:middle line:84% So this is just graphically illustrating 00:42:56.720 --> 00:42:58.160 align:middle line:90% what I just showed you-- 00:42:58.160 --> 00:43:01.400 align:middle line:84% or told you in the first sentence. 00:43:01.400 --> 00:43:03.470 align:middle line:84% There needs to be a constituent, the elephant. 00:43:03.470 --> 00:43:05.120 align:middle line:84% And it's-- yeah, the blue constituent. 00:43:05.120 --> 00:43:07.520 align:middle line:84% It's the one in the tree on the right. 00:43:07.520 --> 00:43:09.445 align:middle line:84% And in the second sentence, there 00:43:09.445 --> 00:43:10.820 align:middle line:84% needs to be a green constituent-- 00:43:10.820 --> 00:43:12.612 align:middle line:84% a constituent "the elephant in my pajamas," 00:43:12.612 --> 00:43:14.660 align:middle line:84% and it's the constituent in the tree on the left. 00:43:14.660 --> 00:43:18.510 align:middle line:90% 00:43:18.510 --> 00:43:22.080 align:middle line:84% Notice that this also makes predictions 00:43:22.080 --> 00:43:23.620 align:middle line:90% about other kinds of sentences. 00:43:23.620 --> 00:43:29.670 align:middle line:84% So if you had thought these trees-- 00:43:29.670 --> 00:43:30.865 align:middle line:90% all this stuff about merge-- 00:43:30.865 --> 00:43:32.490 align:middle line:84% the order in which you'd merge things-- 00:43:32.490 --> 00:43:34.240 align:middle line:84% you don't really need any of this. 00:43:34.240 --> 00:43:37.230 align:middle line:84% Here's an imaginable thing you could think. 00:43:37.230 --> 00:43:40.680 align:middle line:84% You would need to get us to move fairly quickly 00:43:40.680 --> 00:43:41.820 align:middle line:90% through that last slide. 00:43:41.820 --> 00:43:44.530 align:middle line:84% But forgetting about topicalization for a second, 00:43:44.530 --> 00:43:46.530 align:middle line:84% if I told you, "Hey, look, I shot an elephant 00:43:46.530 --> 00:43:48.540 align:middle line:84% in my pajamas," it's ambiguous as to who's 00:43:48.540 --> 00:43:50.040 align:middle line:90% wearing the pajamas. 00:43:50.040 --> 00:43:52.645 align:middle line:84% And I propose to deal with it with these trees, 00:43:52.645 --> 00:43:54.270 align:middle line:84% you can imagine being someone who says, 00:43:54.270 --> 00:43:57.540 align:middle line:84% no, look, "in my pajamas" is just vague. 00:43:57.540 --> 00:43:59.400 align:middle line:84% Somebody's got to be wearing the pajamas, 00:43:59.400 --> 00:44:03.240 align:middle line:84% and you figure it out from context 00:44:03.240 --> 00:44:05.220 align:middle line:84% whether it's me or the elephant, and that's it. 00:44:05.220 --> 00:44:07.650 align:middle line:90% There's no need for any trees. 00:44:07.650 --> 00:44:09.570 align:middle line:84% Just so we're clear, this is not the right way 00:44:09.570 --> 00:44:11.195 align:middle line:84% to think about this, but it's something 00:44:11.195 --> 00:44:13.260 align:middle line:90% you could imagine thinking. 00:44:13.260 --> 00:44:16.140 align:middle line:84% Here-- so the stuff we talked about in the last slide 00:44:16.140 --> 00:44:20.160 align:middle line:84% is one reason not to think about things that way. 00:44:20.160 --> 00:44:22.890 align:middle line:84% Constituency tests seem to nail down 00:44:22.890 --> 00:44:24.870 align:middle line:84% where the pajamas are in the structure, 00:44:24.870 --> 00:44:27.480 align:middle line:84% and that forces you to one reading or another. 00:44:27.480 --> 00:44:30.630 align:middle line:84% Here's another kind of sentence that also suggests-- 00:44:30.630 --> 00:44:33.310 align:middle line:84% where we also make a prediction, which I think is correct. 00:44:33.310 --> 00:44:37.200 align:middle line:84% So if I say "I shot the elephant in my pajamas in a tuxedo"-- 00:44:37.200 --> 00:44:39.750 align:middle line:90% 00:44:39.750 --> 00:44:42.480 align:middle line:90% I think that's grammatical-- 00:44:42.480 --> 00:44:44.970 align:middle line:84% and I think it's a sentence on which somebody is wearing 00:44:44.970 --> 00:44:48.360 align:middle line:84% pajamas and somebody is wearing a tuxedo, 00:44:48.360 --> 00:44:52.010 align:middle line:90% who's wearing the pajamas? 00:44:52.010 --> 00:44:53.690 align:middle line:84% The elephant's wearing the pajamas 00:44:53.690 --> 00:44:55.895 align:middle line:90% and I am wearing the tuxedo. 00:44:55.895 --> 00:44:57.395 align:middle line:84% Does anybody not have that reading-- 00:44:57.395 --> 00:45:01.610 align:middle line:84% or have instead the reading where I am wearing the pajamas 00:45:01.610 --> 00:45:04.760 align:middle line:84% and the elephant is wearing a tuxedo? 00:45:04.760 --> 00:45:08.810 align:middle line:84% Please put a cast out of your mind-- hold on-- 00:45:08.810 --> 00:45:10.310 align:middle line:84% readings in which the pajamas are 00:45:10.310 --> 00:45:11.780 align:middle line:90% inside the tuxedo or something. 00:45:11.780 --> 00:45:12.950 align:middle line:90% [LAUGHTER] 00:45:12.950 --> 00:45:15.320 align:middle line:84% I can see all of you thinking of alternatives, 00:45:15.320 --> 00:45:19.010 align:middle line:84% but just try to imagine cases where one person-- one 00:45:19.010 --> 00:45:21.440 align:middle line:84% entity is wearing pajamas and another one 00:45:21.440 --> 00:45:23.532 align:middle line:90% is wearing the tuxedo. 00:45:23.532 --> 00:45:25.490 align:middle line:84% If that's what's happening-- yes, [INAUDIBLE]?? 00:45:25.490 --> 00:45:26.407 align:middle line:90% AUDIENCE: [INAUDIBLE]. 00:45:26.407 --> 00:45:28.880 align:middle line:84% NORVIN RICHARDS: No, the pajamas are not the tuxedo. 00:45:28.880 --> 00:45:30.575 align:middle line:90% [LAUGHTER] 00:45:30.575 --> 00:45:32.360 align:middle line:90% Ah! 00:45:32.360 --> 00:45:35.330 align:middle line:90% MIT undergrads-- man. 00:45:35.330 --> 00:45:38.912 align:middle line:84% No, there are two sets of clothing and two entities, 00:45:38.912 --> 00:45:41.120 align:middle line:84% and I am also not an elephant just in case any of you 00:45:41.120 --> 00:45:43.700 align:middle line:90% were thinking of that. 00:45:43.700 --> 00:45:46.970 align:middle line:84% And the clothes are not nested, or mixed, 00:45:46.970 --> 00:45:47.910 align:middle line:90% or anything like that. 00:45:47.910 --> 00:45:50.030 align:middle line:84% There is a set of pajamas and there is a tuxedo, 00:45:50.030 --> 00:45:51.110 align:middle line:84% and somebody is wearing the pajamas, 00:45:51.110 --> 00:45:52.190 align:middle line:90% somebody is wearing the tuxedo. 00:45:52.190 --> 00:45:52.970 align:middle line:90% Who's in what? 00:45:52.970 --> 00:45:55.100 align:middle line:84% I just asserted the elephant is in the pajamas 00:45:55.100 --> 00:45:56.790 align:middle line:90% and I'm in the tuxedo. 00:45:56.790 --> 00:45:59.920 align:middle line:84% It can't be the other way around. 00:45:59.920 --> 00:46:01.270 align:middle line:90% And we can see why. 00:46:01.270 --> 00:46:02.345 align:middle line:90% If I was smart enough-- 00:46:02.345 --> 00:46:03.970 align:middle line:84% we'll see if I was smart enough to draw 00:46:03.970 --> 00:46:05.012 align:middle line:90% a tree in the next slide. 00:46:05.012 --> 00:46:05.680 align:middle line:90% Yes, I was! 00:46:05.680 --> 00:46:08.050 align:middle line:90% Hot dog. 00:46:08.050 --> 00:46:08.728 align:middle line:90% Why? 00:46:08.728 --> 00:46:11.020 align:middle line:84% Well, because if you're going to have two prepositional 00:46:11.020 --> 00:46:13.870 align:middle line:84% phrases-- "in my pajamas" and "in a tuxedo"-- 00:46:13.870 --> 00:46:16.300 align:middle line:84% and you're going to attach them in the ways 00:46:16.300 --> 00:46:18.910 align:middle line:84% that we've been attaching prepositional phrases before, 00:46:18.910 --> 00:46:24.000 align:middle line:84% this is the only way to do it to get this order. 00:46:24.000 --> 00:46:27.490 align:middle line:84% Unless you're going to cross lines or something like that, 00:46:27.490 --> 00:46:29.880 align:middle line:84% this is the only way to get this order of words. 00:46:29.880 --> 00:46:33.660 align:middle line:84% So "in my pajamas" has to modify "elephant" and "in the tuxedo" 00:46:33.660 --> 00:46:36.120 align:middle line:84% has to modify the verb phrase, and that's the only reading 00:46:36.120 --> 00:46:38.840 align:middle line:90% this, in fact, has. 00:46:38.840 --> 00:46:43.070 align:middle line:84% So day two of syntax and we're already successfully predicting 00:46:43.070 --> 00:46:46.070 align:middle line:84% facts about reasonably complicated sentences 00:46:46.070 --> 00:46:48.755 align:middle line:84% involving too much clothing in inappropriate places. 00:46:48.755 --> 00:46:51.860 align:middle line:90% 00:46:51.860 --> 00:46:52.370 align:middle line:90% Yeah? 00:46:52.370 --> 00:46:54.537 align:middle line:84% AUDIENCE: We haven't really discussed this directly, 00:46:54.537 --> 00:46:57.007 align:middle line:84% but is it a rule that you can't cross lines in a tree? 00:46:57.007 --> 00:46:59.340 align:middle line:84% NORVIN RICHARDS: I hereby declare that it is a rule that 00:46:59.340 --> 00:47:00.870 align:middle line:90% you cannot cross lines in trees. 00:47:00.870 --> 00:47:04.183 align:middle line:84% Yes, you're right, we have not discussed that. 00:47:04.183 --> 00:47:06.600 align:middle line:84% Back in the day, I would have called this the Ghostbusters 00:47:06.600 --> 00:47:08.275 align:middle line:84% rule just because the Ghostbusters had 00:47:08.275 --> 00:47:09.900 align:middle line:84% a rule that you couldn't cross streams. 00:47:09.900 --> 00:47:10.400 align:middle line:90% Never mind. 00:47:10.400 --> 00:47:12.930 align:middle line:90% 00:47:12.930 --> 00:47:14.220 align:middle line:90% Yeah-- other questions? 00:47:14.220 --> 00:47:17.770 align:middle line:90% 00:47:17.770 --> 00:47:20.890 align:middle line:84% Now, I have this vague sense that some of you 00:47:20.890 --> 00:47:23.350 align:middle line:84% would like to see what the rest of the sentence looks like. 00:47:23.350 --> 00:47:24.058 align:middle line:90% Call it a hunch-- 00:47:24.058 --> 00:47:24.758 align:middle line:90% I don't know. 00:47:24.758 --> 00:47:26.800 align:middle line:84% When you've been an instructor as long as I have, 00:47:26.800 --> 00:47:29.920 align:middle line:84% you get these vague hints like the fact that 12 of you have 00:47:29.920 --> 00:47:31.838 align:middle line:84% asked me where is the subject going-- 00:47:31.838 --> 00:47:33.880 align:middle line:84% where should I put "will"-- how do we do the rest 00:47:33.880 --> 00:47:36.760 align:middle line:84% of the sentence-- why, oh, why are you only showing us verb 00:47:36.760 --> 00:47:38.360 align:middle line:84% phrases and not entire sentences? 00:47:38.360 --> 00:47:41.173 align:middle line:84% So let's do the rest of the sentence. 00:47:41.173 --> 00:47:43.090 align:middle line:84% To do the rest of the sentence-- first of all, 00:47:43.090 --> 00:47:46.180 align:middle line:84% "the girl" is a noun phrase that's kind of boring, 00:47:46.180 --> 00:47:48.340 align:middle line:84% but "girl" is like "the book" or "the garage." 00:47:48.340 --> 00:47:53.980 align:middle line:84% And for "will," we are going to invent a new kind of node. 00:47:53.980 --> 00:47:55.880 align:middle line:84% We're going to call it T for tense. 00:47:55.880 --> 00:47:59.540 align:middle line:90% 00:47:59.540 --> 00:48:03.163 align:middle line:84% Just to prepare you for disappointment, 00:48:03.163 --> 00:48:04.580 align:middle line:84% there are all kinds of things that 00:48:04.580 --> 00:48:08.450 align:middle line:84% can go under T. You can also say "The girl can find 00:48:08.450 --> 00:48:11.060 align:middle line:84% the book in the garage," or "The girl might 00:48:11.060 --> 00:48:13.550 align:middle line:90% find the book in the garage." 00:48:13.550 --> 00:48:15.212 align:middle line:90% I can't talk today. 00:48:15.212 --> 00:48:16.670 align:middle line:84% So there are variety of things that 00:48:16.670 --> 00:48:21.860 align:middle line:84% can go under T, not all of which it's plausible to think 00:48:21.860 --> 00:48:23.390 align:middle line:90% of as instances of tense-- 00:48:23.390 --> 00:48:26.630 align:middle line:84% like "might" is probably not a tense. 00:48:26.630 --> 00:48:28.718 align:middle line:84% If you look it up in traditional grammar, 00:48:28.718 --> 00:48:30.260 align:middle line:84% they're not going to call it a tense, 00:48:30.260 --> 00:48:31.430 align:middle line:84% but we're going to call it a tense. 00:48:31.430 --> 00:48:33.140 align:middle line:84% We basically just need a name for it. 00:48:33.140 --> 00:48:35.870 align:middle line:90% So there's this thing. 00:48:35.870 --> 00:48:42.380 align:middle line:84% Notice it's not a verb, so you can't say things like the girl 00:48:42.380 --> 00:48:45.110 align:middle line:90% will the book. 00:48:45.110 --> 00:48:47.240 align:middle line:84% It's something else, and our name for it 00:48:47.240 --> 00:48:48.290 align:middle line:90% is going to be "tense." 00:48:48.290 --> 00:48:49.520 align:middle line:84% People have called it other things. 00:48:49.520 --> 00:48:51.353 align:middle line:84% They've called it the auxiliary or whatever, 00:48:51.353 --> 00:48:54.150 align:middle line:84% but tense is a conventional name for it in the literature, 00:48:54.150 --> 00:48:55.760 align:middle line:90% so I'm going to call it that. 00:48:55.760 --> 00:48:56.630 align:middle line:90% Yeah? 00:48:56.630 --> 00:48:59.870 align:middle line:84% AUDIENCE: So how would you differentiate the tense 00:48:59.870 --> 00:49:01.938 align:middle line:90% from [INAUDIBLE]? 00:49:01.938 --> 00:49:03.355 align:middle line:84% NORVIN RICHARDS: Well-- so those-- 00:49:03.355 --> 00:49:04.855 align:middle line:84% AUDIENCE: "The girl quickly found--" 00:49:04.855 --> 00:49:07.230 align:middle line:84% NORVIN RICHARDS: "Quickly found the book in the garage"-- 00:49:07.230 --> 00:49:08.320 align:middle line:90% yeah, good question. 00:49:08.320 --> 00:49:10.620 align:middle line:90% Let me give you the-- 00:49:10.620 --> 00:49:17.040 align:middle line:84% let me give you an answer that I won't be able to explain soon, 00:49:17.040 --> 00:49:19.740 align:middle line:90% but that's a very nice question. 00:49:19.740 --> 00:49:21.460 align:middle line:90% So the girl will find it-- 00:49:21.460 --> 00:49:22.710 align:middle line:90% I'll just say that-- 00:49:22.710 --> 00:49:31.080 align:middle line:90% and the girl quickly found it. 00:49:31.080 --> 00:49:33.600 align:middle line:84% And you want to know what's the difference between these two 00:49:33.600 --> 00:49:35.063 align:middle line:90% things. 00:49:35.063 --> 00:49:35.980 align:middle line:90% Here's one difference. 00:49:35.980 --> 00:49:40.290 align:middle line:84% There are a bunch of phenomena that this gets to participate 00:49:40.290 --> 00:49:41.370 align:middle line:90% in this doesn't. 00:49:41.370 --> 00:49:45.720 align:middle line:84% So for example, if I want to make this into a question, 00:49:45.720 --> 00:49:52.643 align:middle line:84% I can ask the question, "Will the girl find it?" 00:49:52.643 --> 00:49:54.810 align:middle line:84% where I put "will" at the beginning of the sentence, 00:49:54.810 --> 00:49:57.692 align:middle line:84% and that makes the sentence a question. 00:49:57.692 --> 00:50:00.150 align:middle line:84% You can put "quickly" at the beginning of the sentence too, 00:50:00.150 --> 00:50:01.320 align:middle line:90% I guess-- 00:50:01.320 --> 00:50:05.160 align:middle line:90% "Quickly, the girl found it." 00:50:05.160 --> 00:50:08.870 align:middle line:84% That's a grammatical sentence, but it's not a question. 00:50:08.870 --> 00:50:11.190 align:middle line:90% It doesn't make it a question. 00:50:11.190 --> 00:50:14.970 align:middle line:84% So that gives us some reason to want these 00:50:14.970 --> 00:50:17.430 align:middle line:90% to have different statuses. 00:50:17.430 --> 00:50:19.350 align:middle line:84% I'll give you another reason to have 00:50:19.350 --> 00:50:22.140 align:middle line:90% them have different statuses. 00:50:22.140 --> 00:50:24.860 align:middle line:90% If I ask you-- 00:50:24.860 --> 00:50:27.050 align:middle line:90% well, is this true? 00:50:27.050 --> 00:50:28.010 align:middle line:90% Maybe I'm about to lie. 00:50:28.010 --> 00:50:29.900 align:middle line:84% Maybe I should quit while I'm ahead. 00:50:29.900 --> 00:50:30.800 align:middle line:90% I think this is true. 00:50:30.800 --> 00:50:33.290 align:middle line:84% If I ask you, "Who's going to find the book?" 00:50:33.290 --> 00:50:36.500 align:middle line:90% you can answer "The girl will." 00:50:36.500 --> 00:50:38.970 align:middle line:84% But if I ask you "Who found the book?" 00:50:38.970 --> 00:50:41.830 align:middle line:84% it would be odd for you to say "The girl quickly." 00:50:41.830 --> 00:50:43.730 align:middle line:90% That's-- I think-- is that true? 00:50:43.730 --> 00:50:45.640 align:middle line:90% Did I just lie? 00:50:45.640 --> 00:50:46.750 align:middle line:90% I think that's true. 00:50:46.750 --> 00:50:48.250 align:middle line:84% So that's another reason to think 00:50:48.250 --> 00:50:49.810 align:middle line:84% there's some difference between "will" and "quickly." 00:50:49.810 --> 00:50:51.370 align:middle line:90% Now, what's the difference? 00:50:51.370 --> 00:50:53.548 align:middle line:84% Well, we'll get to that, but-- or like how long 00:50:53.548 --> 00:50:55.090 align:middle line:84% are we going to explain these things, 00:50:55.090 --> 00:50:56.332 align:middle line:90% and we'll get to that too. 00:50:56.332 --> 00:50:58.540 align:middle line:84% This is just meant to convince you that maybe they're 00:50:58.540 --> 00:50:59.510 align:middle line:90% different things-- 00:50:59.510 --> 00:51:02.270 align:middle line:90% they have different properties. 00:51:02.270 --> 00:51:04.130 align:middle line:90% OK, cool. 00:51:04.130 --> 00:51:05.330 align:middle line:90% So far, so good. 00:51:05.330 --> 00:51:07.790 align:middle line:84% Now, we need to do some more merging. 00:51:07.790 --> 00:51:09.650 align:middle line:90% Here's what we'll do. 00:51:09.650 --> 00:51:12.140 align:middle line:84% We'll merge T with the verb phrase, 00:51:12.140 --> 00:51:15.510 align:middle line:84% and then we'll merge "the girl" with the result 00:51:15.510 --> 00:51:18.440 align:middle line:84% so that the entire sentence will be a TP-- 00:51:18.440 --> 00:51:20.720 align:middle line:90% a tense phrase. 00:51:20.720 --> 00:51:22.070 align:middle line:90% Everything is a phrase. 00:51:22.070 --> 00:51:24.027 align:middle line:84% There are these, there are noun phrases, 00:51:24.027 --> 00:51:26.360 align:middle line:84% there are prepositional phrases, there are verb phrases. 00:51:26.360 --> 00:51:28.693 align:middle line:84% And now, we have a new kind of phrase-- a tense phrase-- 00:51:28.693 --> 00:51:32.450 align:middle line:84% which is the result of merging tense will 00:51:32.450 --> 00:51:34.640 align:middle line:90% with a bunch of things. 00:51:34.640 --> 00:51:39.560 align:middle line:84% There-- that's how we'll do sentences Cool. 00:51:39.560 --> 00:51:41.655 align:middle line:90% That was easy and painless. 00:51:41.655 --> 00:51:42.155 align:middle line:90% Yeah? 00:51:42.155 --> 00:51:44.840 align:middle line:84% AUDIENCE: What does topicalize mean again exactly? 00:51:44.840 --> 00:51:47.420 align:middle line:84% NORVIN RICHARDS: Oh, it's the name 00:51:47.420 --> 00:51:51.050 align:middle line:84% that we've got for taking constituents, putting them 00:51:51.050 --> 00:51:53.720 align:middle line:84% at the beginning of the sentence, 00:51:53.720 --> 00:51:58.310 align:middle line:84% and it adds some kind of oomph to the thing 00:51:58.310 --> 00:52:00.950 align:middle line:84% that you put at the beginning of the sentence. 00:52:00.950 --> 00:52:02.870 align:middle line:84% So if I say things like "in the garage, 00:52:02.870 --> 00:52:06.800 align:middle line:84% the girl will find the book," that's a sentence of English. 00:52:06.800 --> 00:52:09.020 align:middle line:84% It's not the most natural way to say it-- 00:52:09.020 --> 00:52:11.270 align:middle line:84% you would only say it if you wanted 00:52:11.270 --> 00:52:14.540 align:middle line:84% to put some kind of special emphasis on the garage. 00:52:14.540 --> 00:52:18.060 align:middle line:84% And we will not ever, in this class, 00:52:18.060 --> 00:52:19.520 align:middle line:84% try to be more definite than that 00:52:19.520 --> 00:52:21.020 align:middle line:84% about the meaning of topicalization. 00:52:21.020 --> 00:52:25.790 align:middle line:84% But it's a handy phenomenon for finding constituents. 00:52:25.790 --> 00:52:27.190 align:middle line:90% Yeah? 00:52:27.190 --> 00:52:31.510 align:middle line:84% AUDIENCE: Every node on the tree is a constituent? 00:52:31.510 --> 00:52:33.850 align:middle line:90% NORVIN RICHARDS: Yes. 00:52:33.850 --> 00:52:36.970 align:middle line:84% AUDIENCE: What is "will find the book in the garage"? 00:52:36.970 --> 00:52:38.710 align:middle line:90% NORVIN RICHARDS: It's a T bar. 00:52:38.710 --> 00:52:41.620 align:middle line:84% AUDIENCE: Well, yes, but we have constituency tests 00:52:41.620 --> 00:52:44.890 align:middle line:84% that we use to tell whether something is a constituent. 00:52:44.890 --> 00:52:48.522 align:middle line:84% Does that work for a tense phrase? 00:52:48.522 --> 00:52:49.480 align:middle line:90% NORVIN RICHARDS: Yeah-- 00:52:49.480 --> 00:52:50.605 align:middle line:90% OK, that's a good question. 00:52:50.605 --> 00:52:52.462 align:middle line:90% [LAUGHTER] 00:52:52.462 --> 00:52:53.420 align:middle line:90% That's a good question. 00:52:53.420 --> 00:52:53.920 align:middle line:90% Let's see. 00:52:53.920 --> 00:52:56.860 align:middle line:90% So let me-- so let's see. 00:52:56.860 --> 00:52:58.580 align:middle line:90% You would like to know-- 00:52:58.580 --> 00:52:59.830 align:middle line:90% is this what you're asking me? 00:52:59.830 --> 00:53:02.320 align:middle line:90% Let's see. 00:53:02.320 --> 00:53:03.940 align:middle line:90% We've got three things-- 00:53:03.940 --> 00:53:06.880 align:middle line:90% 00:53:06.880 --> 00:53:14.900 align:middle line:84% the noun phrase, "the girl," and the T "will," 00:53:14.900 --> 00:53:18.560 align:middle line:84% and the verb phrase find "the book in the garage." 00:53:18.560 --> 00:53:19.120 align:middle line:90% Oh, here. 00:53:19.120 --> 00:53:27.280 align:middle line:90% 00:53:27.280 --> 00:53:29.720 align:middle line:84% We have a technical term for this. 00:53:29.720 --> 00:53:32.830 align:middle line:90% It's called a triangle. 00:53:32.830 --> 00:53:35.260 align:middle line:84% It's used when you do not wish to draw 00:53:35.260 --> 00:53:37.270 align:middle line:90% the inside of a structure. 00:53:37.270 --> 00:53:40.270 align:middle line:84% So this is just an abbreviation for that. 00:53:40.270 --> 00:53:41.230 align:middle line:90% Is that OK? 00:53:41.230 --> 00:53:42.730 align:middle line:84% So I've just used a triangle here so 00:53:42.730 --> 00:53:44.605 align:middle line:84% that I won't have to draw that entire phrase. 00:53:44.605 --> 00:53:46.990 align:middle line:84% What I really need is like a stamp or something 00:53:46.990 --> 00:53:50.230 align:middle line:84% that I can use on the blackboard that will create these phrase 00:53:50.230 --> 00:53:52.120 align:middle line:90% structures for me. 00:53:52.120 --> 00:53:54.160 align:middle line:84% We'll tell you on problem sets if we 00:53:54.160 --> 00:53:55.525 align:middle line:90% want you to not use triangles. 00:53:55.525 --> 00:53:57.400 align:middle line:84% Sometimes, we'll ask you not to use triangles 00:53:57.400 --> 00:53:59.590 align:middle line:84% so that we can make you show us everything 00:53:59.590 --> 00:54:01.760 align:middle line:84% that you think about the insides of structures, 00:54:01.760 --> 00:54:03.920 align:middle line:84% but for now, I'll use a triangle. 00:54:03.920 --> 00:54:05.452 align:middle line:84% So I just casually said, hey, I've 00:54:05.452 --> 00:54:07.660 align:middle line:84% got a good idea-- we'll merge "will" first with this, 00:54:07.660 --> 00:54:10.623 align:middle line:84% and then we'll merge the result with that. 00:54:10.623 --> 00:54:12.790 align:middle line:84% If we're doing merge the way we've been doing merge, 00:54:12.790 --> 00:54:14.740 align:middle line:84% there's only one other option, which is this. 00:54:14.740 --> 00:54:20.770 align:middle line:90% 00:54:20.770 --> 00:54:24.580 align:middle line:84% And so you would like to know, how do we know 00:54:24.580 --> 00:54:28.180 align:middle line:90% that it's what's on the slide? 00:54:28.180 --> 00:54:29.590 align:middle line:84% How do we know that it's not what 00:54:29.590 --> 00:54:31.570 align:middle line:90% I've got here on the board? 00:54:31.570 --> 00:54:39.670 align:middle line:84% Just so nobody gets confused, this is not the right tree. 00:54:39.670 --> 00:54:41.500 align:middle line:84% That's the right tree, but now let 00:54:41.500 --> 00:54:44.140 align:middle line:84% me see if I can come up with anything that will convince you 00:54:44.140 --> 00:54:46.480 align:middle line:90% that it's the right tree. 00:54:46.480 --> 00:54:53.840 align:middle line:84% First of all, if we tried to use topicalization, which is 00:54:53.840 --> 00:54:57.300 align:middle line:90% the test that we had before-- 00:54:57.300 --> 00:54:59.750 align:middle line:84% if we tried to use it on the tree here and we tried 00:54:59.750 --> 00:55:03.020 align:middle line:84% to topicalize T-bar, we'd end up with-- 00:55:03.020 --> 00:55:05.360 align:middle line:84% "I said the girl would find the book in the garage 00:55:05.360 --> 00:55:10.200 align:middle line:84% and will find the book in the garage, the girl." 00:55:10.200 --> 00:55:12.960 align:middle line:84% Now, as I mentioned before, my grasp of English 00:55:12.960 --> 00:55:16.210 align:middle line:84% is now a little shaky after years of abuse, 00:55:16.210 --> 00:55:18.840 align:middle line:90% but I think that's not great. 00:55:18.840 --> 00:55:21.180 align:middle line:90% Do people agree? 00:55:21.180 --> 00:55:24.030 align:middle line:84% So if there is a constituent T-bar here-- 00:55:24.030 --> 00:55:26.998 align:middle line:84% I tried to soften you up for this possibility earlier-- 00:55:26.998 --> 00:55:29.040 align:middle line:84% we need it to be the kind of constituent to which 00:55:29.040 --> 00:55:30.510 align:middle line:90% topicalization can't apply. 00:55:30.510 --> 00:55:33.430 align:middle line:84% That is, topicalization doesn't get to apply to everything. 00:55:33.430 --> 00:55:35.670 align:middle line:84% And I think I also floated the possibility 00:55:35.670 --> 00:55:37.290 align:middle line:84% that it cares about whether you're 00:55:37.290 --> 00:55:40.260 align:middle line:90% looking at a P or a bar-- 00:55:40.260 --> 00:55:44.580 align:middle line:84% that you can't topicalize bars, you can only topicalize P's. 00:55:44.580 --> 00:55:46.260 align:middle line:84% This might be another case where we 00:55:46.260 --> 00:55:48.660 align:middle line:84% want to take that possibility seriously. 00:55:48.660 --> 00:55:50.370 align:middle line:90% Yeah? 00:55:50.370 --> 00:55:55.340 align:middle line:84% AUDIENCE: It seems like that-- the one on the board 00:55:55.340 --> 00:55:57.090 align:middle line:84% works better as a constituent, because you 00:55:57.090 --> 00:56:00.090 align:middle line:84% could be like, "Who will find the book in the garage?" 00:56:00.090 --> 00:56:02.430 align:middle line:90% and then say "The girl will." 00:56:02.430 --> 00:56:05.382 align:middle line:84% But if you say "What will the girl do," you can't say 00:56:05.382 --> 00:56:06.840 align:middle line:84% "Will find the book in the garage." 00:56:06.840 --> 00:56:08.340 align:middle line:84% NORVIN RICHARDS: You can't say "will 00:56:08.340 --> 00:56:09.840 align:middle line:90% find the book in the garage"-- 00:56:09.840 --> 00:56:15.120 align:middle line:84% if-- I am astonished by what you have just said. 00:56:15.120 --> 00:56:17.190 align:middle line:84% If I've just-- you've just said "The girl will 00:56:17.190 --> 00:56:18.480 align:middle line:90% find the book in the garage." 00:56:18.480 --> 00:56:23.230 align:middle line:84% Can I say "Will find the book in the garage?" 00:56:23.230 --> 00:56:25.990 align:middle line:90% Some think yes, others think no. 00:56:25.990 --> 00:56:31.190 align:middle line:84% We'll have a wrestling match outside later. 00:56:31.190 --> 00:56:32.752 align:middle line:90% AUDIENCE: That makes sense. 00:56:32.752 --> 00:56:34.210 align:middle line:84% But if you're astonished, you could 00:56:34.210 --> 00:56:36.252 align:middle line:84% say "The girl will find the book in the garage"-- 00:56:36.252 --> 00:56:37.310 align:middle line:90% "The girl will?" 00:56:37.310 --> 00:56:39.435 align:middle line:84% NORVIN RICHARDS: (QUESTIONINGLY) "The girl will?"-- 00:56:39.435 --> 00:56:42.730 align:middle line:90% ah, yeah, that's true. 00:56:42.730 --> 00:56:46.945 align:middle line:84% You guys are discovering a phenomenon called ellipsis. 00:56:46.945 --> 00:56:50.110 align:middle line:90% 00:56:50.110 --> 00:56:53.800 align:middle line:84% Ellipsis is a phenomenon whereby you 00:56:53.800 --> 00:56:57.430 align:middle line:84% can take chunks of the sentence and fail to pronounce them. 00:56:57.430 --> 00:57:01.910 align:middle line:84% You can remove chunks of the sentence. 00:57:01.910 --> 00:57:04.480 align:middle line:84% So if I ask you a question like, "Who will 00:57:04.480 --> 00:57:06.280 align:middle line:90% find the book in the garage?" 00:57:06.280 --> 00:57:08.860 align:middle line:90% you can say "The girl will." 00:57:08.860 --> 00:57:10.840 align:middle line:84% Or I can say things like "The boys will 00:57:10.840 --> 00:57:13.225 align:middle line:84% find the book in the garage and the girl will too." 00:57:13.225 --> 00:57:16.090 align:middle line:90% 00:57:16.090 --> 00:57:20.105 align:middle line:84% All of these-- the way we would talk about them 00:57:20.105 --> 00:57:22.480 align:middle line:84% in terms of tree, like the one that's on the slide there, 00:57:22.480 --> 00:57:24.340 align:middle line:84% is in terms of what's called VP ellipses. 00:57:24.340 --> 00:57:24.840 align:middle line:90% It. 00:57:24.840 --> 00:57:28.870 align:middle line:84% Is this process of silencing constituents-- 00:57:28.870 --> 00:57:31.150 align:middle line:90% silencing phrases-- 00:57:31.150 --> 00:57:33.520 align:middle line:84% has applied to the verb phrase when 00:57:33.520 --> 00:57:43.810 align:middle line:84% you say things like "Who will find the book, (blah, blah, 00:57:43.810 --> 00:57:45.070 align:middle line:90% blah)?"-- 00:57:45.070 --> 00:57:45.790 align:middle line:90% "The girl will." 00:57:45.790 --> 00:57:48.590 align:middle line:90% 00:57:48.590 --> 00:57:50.290 align:middle line:84% So there's a blank here, and the blank 00:57:50.290 --> 00:57:56.080 align:middle line:84% is understood as being the same as the preceding verb phrase. 00:57:56.080 --> 00:58:00.098 align:middle line:84% And VP ellipsis is a huge phenomenon, 00:58:00.098 --> 00:58:02.140 align:middle line:84% which we will get a chance to talk about actually 00:58:02.140 --> 00:58:06.070 align:middle line:84% because it's going to be useful in a couple of weeks 00:58:06.070 --> 00:58:07.630 align:middle line:90% for other phenomena. 00:58:07.630 --> 00:58:10.330 align:middle line:90% 00:58:10.330 --> 00:58:13.300 align:middle line:84% Notice-- I'm trying to think if there's 00:58:13.300 --> 00:58:15.850 align:middle line:90% a way that I can convince you. 00:58:15.850 --> 00:58:18.250 align:middle line:84% And so that's not where you were going with this. 00:58:18.250 --> 00:58:20.980 align:middle line:84% Where you were going with it was, hey, look, "The girl will" 00:58:20.980 --> 00:58:21.748 align:middle line:90% is a constituent. 00:58:21.748 --> 00:58:23.290 align:middle line:84% There is a thing that we want to have 00:58:23.290 --> 00:58:26.770 align:middle line:84% be a constituent-- we want it to be a T-bar like here, 00:58:26.770 --> 00:58:29.230 align:middle line:84% whereas what I'm saying is, no, we 00:58:29.230 --> 00:58:31.492 align:middle line:84% want the structure on the slide, and there 00:58:31.492 --> 00:58:32.950 align:middle line:84% is such a thing as VP ellipsis that 00:58:32.950 --> 00:58:37.930 align:middle line:84% allows you to fail to pronounce verb phrases. 00:58:37.930 --> 00:58:43.610 align:middle line:84% Let me see if I can come up with another test that 00:58:43.610 --> 00:58:48.985 align:middle line:84% will help you believe in one of these over there. 00:58:48.985 --> 00:58:51.560 align:middle line:84% There are other phenomena that people 00:58:51.560 --> 00:58:54.402 align:middle line:84% use to diagnose constituent structure, 00:58:54.402 --> 00:58:56.360 align:middle line:84% and I'm trying to think if there are any that I 00:58:56.360 --> 00:59:00.200 align:middle line:90% can introduce painlessly. 00:59:00.200 --> 00:59:02.240 align:middle line:90% You-- ah, no. 00:59:02.240 --> 00:59:03.987 align:middle line:84% So I was about to introduce coordination, 00:59:03.987 --> 00:59:04.820 align:middle line:90% but I think I won't. 00:59:04.820 --> 00:59:05.403 align:middle line:90% You know what? 00:59:05.403 --> 00:59:07.610 align:middle line:84% Your TAs will prove to you that it's 00:59:07.610 --> 00:59:09.830 align:middle line:84% the tree on the left and not the tree on the right 00:59:09.830 --> 00:59:13.900 align:middle line:90% in the recitation section. 00:59:13.900 --> 00:59:15.650 align:middle line:84% So definitely go to the recitation session 00:59:15.650 --> 00:59:18.800 align:middle line:84% this week, because your TAs will help you with this. 00:59:18.800 --> 00:59:23.180 align:middle line:84% I don't want to go too far down this path because there are-- 00:59:23.180 --> 00:59:27.710 align:middle line:84% if I try to show you the constituent tests that 00:59:27.710 --> 00:59:30.570 align:middle line:84% would help us to find that, I think we'll be in trouble. 00:59:30.570 --> 00:59:35.740 align:middle line:84% There is one-- well, maybe I can do this now. 00:59:35.740 --> 00:59:36.820 align:middle line:90% Can I do this now? 00:59:36.820 --> 00:59:40.970 align:middle line:90% 00:59:40.970 --> 00:59:42.950 align:middle line:84% I'm going to get in so much trouble. 00:59:42.950 --> 00:59:45.137 align:middle line:90% This is not going to go well. 00:59:45.137 --> 00:59:47.720 align:middle line:84% I'm going to regret doing this, and we're almost out of clock. 00:59:47.720 --> 00:59:49.987 align:middle line:84% Tell you what-- I will prove to you-- 00:59:49.987 --> 00:59:52.070 align:middle line:84% you don't have to-- so you should go to recitation 00:59:52.070 --> 00:59:54.380 align:middle line:84% session, but I will prove to you in the next class 00:59:54.380 --> 00:59:57.140 align:middle line:84% that that's the right tree and this is the wrong tree, 00:59:57.140 --> 00:59:59.940 align:middle line:84% but it's going to require some setup. 00:59:59.940 --> 01:00:01.730 align:middle line:84% And if I try to do the setup now, 01:00:01.730 --> 01:00:04.370 align:middle line:84% we'll be in the middle of the setup at the end of class, 01:00:04.370 --> 01:00:06.080 align:middle line:90% and we'll all go away unhappy. 01:00:06.080 --> 01:00:09.180 align:middle line:84% So this is a really excellent question, 01:00:09.180 --> 01:00:13.790 align:middle line:84% and you're right to be skeptical and alarmed. 01:00:13.790 --> 01:00:19.160 align:middle line:84% But everybody write the tree that's on the board here, 01:00:19.160 --> 01:00:22.280 align:middle line:84% and I will also reproduce it on the next set 01:00:22.280 --> 01:00:24.950 align:middle line:90% of slides for Thursday. 01:00:24.950 --> 01:00:27.260 align:middle line:84% And then you will go into spring break understanding 01:00:27.260 --> 01:00:29.390 align:middle line:84% why I think it's that tree and not this tree. 01:00:29.390 --> 01:00:33.080 align:middle line:84% But if I tried to show it to you now, I'll just hurt all of us. 01:00:33.080 --> 01:00:34.663 align:middle line:90% Yeah good questions. 01:00:34.663 --> 01:00:36.205 align:middle line:84% Are there other questions about this. 01:00:36.205 --> 01:00:38.038 align:middle line:84% I'll very quickly move away from this issue. 01:00:38.038 --> 01:00:40.760 align:middle line:90% 01:00:40.760 --> 01:00:45.020 align:middle line:84% So I want to go back to the verb phrase just as an exercise 01:00:45.020 --> 01:00:48.872 align:middle line:84% in making sure that we all know what we're doing. 01:00:48.872 --> 01:00:50.330 align:middle line:84% So we're going to do "I will tickle 01:00:50.330 --> 01:00:52.100 align:middle line:90% the child with the feather. 01:00:52.100 --> 01:00:55.023 align:middle line:84% Ambiguous sentence-- does everybody agree? 01:00:55.023 --> 01:00:56.690 align:middle line:84% So what are the things that it can mean? 01:00:56.690 --> 01:01:00.550 align:middle line:90% 01:01:00.550 --> 01:01:02.050 align:middle line:84% Can somebody paraphrase them for me? 01:01:02.050 --> 01:01:03.300 align:middle line:90% Joseph? 01:01:03.300 --> 01:01:06.520 align:middle line:84% AUDIENCE: Either there's a child that is holding a feather 01:01:06.520 --> 01:01:09.080 align:middle line:84% and your going to go tickle him or her, 01:01:09.080 --> 01:01:12.755 align:middle line:84% or you're going to go tickle the child, 01:01:12.755 --> 01:01:14.255 align:middle line:84% and you have a feather in your hand, 01:01:14.255 --> 01:01:16.050 align:middle line:84% and you're going to use the feather. 01:01:16.050 --> 01:01:18.180 align:middle line:84% NORVIN RICHARDS: Yep, nicely paraphrased. 01:01:18.180 --> 01:01:21.120 align:middle line:84% So give me neither of those things. 01:01:21.120 --> 01:01:27.690 align:middle line:84% We think that we can build this in a couple of different ways. 01:01:27.690 --> 01:01:29.430 align:middle line:90% So "I will tickle the child"-- 01:01:29.430 --> 01:01:33.930 align:middle line:84% we have two places to attach "with the feather." 01:01:33.930 --> 01:01:35.642 align:middle line:90% So two trees for this. 01:01:35.642 --> 01:01:37.350 align:middle line:84% I'm going through this fast because we're 01:01:37.350 --> 01:01:39.473 align:middle line:84% running low on time, but is this all clear? 01:01:39.473 --> 01:01:40.890 align:middle line:84% This is just the same kind of tree 01:01:40.890 --> 01:01:44.220 align:middle line:84% that we've been drawing up until now. 01:01:44.220 --> 01:01:48.300 align:middle line:84% And constituency tests, like topicalization, 01:01:48.300 --> 01:01:51.720 align:middle line:84% tell us that we're right to associate 01:01:51.720 --> 01:01:54.130 align:middle line:84% these two different trees with two different meanings. 01:01:54.130 --> 01:01:56.965 align:middle line:84% So these trees are basically just like the trees we did-- 01:01:56.965 --> 01:01:58.590 align:middle line:84% in fact, they're identical to the trees 01:01:58.590 --> 01:02:03.120 align:middle line:84% that we did for "I shot an elephant in my pajamas." 01:02:03.120 --> 01:02:06.810 align:middle line:84% The prepositional phrase that's at the end of the sentence 01:02:06.810 --> 01:02:08.640 align:middle line:84% can either be part of the noun phrase 01:02:08.640 --> 01:02:10.223 align:middle line:84% that's also at the end of the sentence 01:02:10.223 --> 01:02:12.570 align:middle line:84% or it can modify the verb phrase. 01:02:12.570 --> 01:02:15.495 align:middle line:84% It depends on whether we're doing, in this case, 01:02:15.495 --> 01:02:17.370 align:middle line:84% the tickling with the feather or whether it's 01:02:17.370 --> 01:02:18.828 align:middle line:84% the child who has a feather and I'm 01:02:18.828 --> 01:02:21.810 align:middle line:84% going to tickle them some other way-- maybe with my fingers. 01:02:21.810 --> 01:02:23.310 align:middle line:84% And then if we do topicalization, 01:02:23.310 --> 01:02:25.110 align:middle line:84% that makes the meanings collapse. 01:02:25.110 --> 01:02:31.140 align:middle line:84% So "The child I will tickle with the feather" only 01:02:31.140 --> 01:02:32.490 align:middle line:90% has the tree on the left. 01:02:32.490 --> 01:02:34.240 align:middle line:84% "The child with the feather I will tickle" 01:02:34.240 --> 01:02:35.990 align:middle line:84% only has the tree on the right and it only 01:02:35.990 --> 01:02:37.270 align:middle line:90% has the meaning on the right. 01:02:37.270 --> 01:02:38.970 align:middle line:90% Does everybody see that? 01:02:38.970 --> 01:02:40.740 align:middle line:84% This is review of where we are so 01:02:40.740 --> 01:02:43.500 align:middle line:84% far with the constituent structures-- 01:02:43.500 --> 01:02:45.400 align:middle line:90% constituency tests that we have. 01:02:45.400 --> 01:02:47.970 align:middle line:84% And as I said, I will reveal another one to you 01:02:47.970 --> 01:02:50.580 align:middle line:84% on Thursday, which will help me to get rid of this tree. 01:02:50.580 --> 01:02:51.360 align:middle line:90% Yeah? 01:02:51.360 --> 01:02:52.440 align:middle line:84% AUDIENCE: Oh, so that second sentence-- 01:02:52.440 --> 01:02:54.750 align:middle line:84% I think this is a feature of English grammar, which 01:02:54.750 --> 01:02:57.545 align:middle line:84% is that you could understand "the child with the feather" 01:02:57.545 --> 01:03:00.720 align:middle line:84% [INAUDIBLE] as a nondestructive modifier. 01:03:00.720 --> 01:03:03.270 align:middle line:84% So that I have a child with a feather and I will tickle-- 01:03:03.270 --> 01:03:05.312 align:middle line:84% NORVIN RICHARDS: "I will tickle people generally. 01:03:05.312 --> 01:03:08.880 align:middle line:84% The child with the feather, I will tickle." 01:03:08.880 --> 01:03:10.230 align:middle line:90% Really? 01:03:10.230 --> 01:03:12.270 align:middle line:90% Maybe. 01:03:12.270 --> 01:03:13.035 align:middle line:90% If so-- 01:03:13.035 --> 01:03:14.220 align:middle line:84% AUDIENCE: "I, the child with the feather"-- 01:03:14.220 --> 01:03:16.050 align:middle line:84% NORVIN RICHARDS: "I, the child with the feather, will tickle." 01:03:16.050 --> 01:03:16.680 align:middle line:90% AUDIENCE: --[INAUDIBLE]. 01:03:16.680 --> 01:03:19.180 align:middle line:84% NORVIN RICHARDS: I think I want "the child with the feather" 01:03:19.180 --> 01:03:20.760 align:middle line:90% to be after "I" for that. 01:03:20.760 --> 01:03:22.410 align:middle line:84% I think that was Kateryna's point. 01:03:22.410 --> 01:03:23.160 align:middle line:90% AUDIENCE: I am the child. 01:03:23.160 --> 01:03:25.243 align:middle line:84% NORVIN RICHARDS: "I, the child with the feather"-- 01:03:25.243 --> 01:03:30.150 align:middle line:84% "I, the professor in this class, demand cookies." 01:03:30.150 --> 01:03:34.450 align:middle line:84% I think that's grammatical-- even plausible. 01:03:34.450 --> 01:03:34.950 align:middle line:90% Cool. 01:03:34.950 --> 01:03:36.750 align:middle line:90% So yeah, good point. 01:03:36.750 --> 01:03:40.975 align:middle line:84% There could be other structures to think about here. 01:03:40.975 --> 01:03:42.350 align:middle line:84% Because we're almost out of time, 01:03:42.350 --> 01:03:45.500 align:middle line:84% let me talk quickly through some terminology, 01:03:45.500 --> 01:03:47.570 align:middle line:90% and then I'll let you guys go. 01:03:47.570 --> 01:03:49.770 align:middle line:90% First bit of terminology-- 01:03:49.770 --> 01:03:52.280 align:middle line:84% when people are talking about trees 01:03:52.280 --> 01:03:55.310 align:middle line:84% they often use feminine kinship terminology 01:03:55.310 --> 01:03:59.310 align:middle line:84% to talk about relationships between nodes in the tree. 01:03:59.310 --> 01:04:06.740 align:middle line:84% So if I want to say something like "I will tickle the child," 01:04:06.740 --> 01:04:11.330 align:middle line:84% we have pairs of nodes that were merged together 01:04:11.330 --> 01:04:12.890 align:middle line:90% to form larger objects. 01:04:12.890 --> 01:04:15.890 align:middle line:84% Like in this tree, we merged the T-- 01:04:15.890 --> 01:04:19.160 align:middle line:84% "will"-- with the verb phrase-- "tickle the child"-- 01:04:19.160 --> 01:04:23.240 align:middle line:84% to form the T-bar that's just above them. 01:04:23.240 --> 01:04:25.820 align:middle line:84% What we say about those is that they are sisters. 01:04:25.820 --> 01:04:28.310 align:middle line:84% So if you have two things that are the two things that 01:04:28.310 --> 01:04:30.500 align:middle line:84% were merged together to create one larger thing, 01:04:30.500 --> 01:04:32.870 align:middle line:90% those two things are sisters. 01:04:32.870 --> 01:04:38.000 align:middle line:84% Similarly, we say that VP is the mother of V and NP. 01:04:38.000 --> 01:04:40.970 align:middle line:84% And we also use the word "daughter"-- 01:04:40.970 --> 01:04:42.577 align:middle line:90% sorry. 01:04:42.577 --> 01:04:43.910 align:middle line:90% We also use the word "daughter." 01:04:43.910 --> 01:04:46.202 align:middle line:84% Although I don't have it on the slide, I should add it. 01:04:46.202 --> 01:04:48.800 align:middle line:84% So you say that the V is the daughter of VP. 01:04:48.800 --> 01:04:50.930 align:middle line:84% That's as far as people go with kinship terms, 01:04:50.930 --> 01:04:53.067 align:middle line:84% so you don't hear about grandmothers, 01:04:53.067 --> 01:04:54.900 align:middle line:84% or aunts, or cousins, or anything like that. 01:04:54.900 --> 01:04:58.130 align:middle line:84% But people use those terms to talk about relationships 01:04:58.130 --> 01:05:00.440 align:middle line:90% between nodes in the tree. 01:05:00.440 --> 01:05:04.460 align:middle line:84% The relationship of motherhood is-- 01:05:04.460 --> 01:05:09.260 align:middle line:84% so the relation that is created between two nodes that 01:05:09.260 --> 01:05:12.380 align:middle line:84% are merged and the new node that you create via merge, 01:05:12.380 --> 01:05:14.060 align:middle line:84% that's the relationship of what's 01:05:14.060 --> 01:05:15.410 align:middle line:90% called immediate domination. 01:05:15.410 --> 01:05:18.500 align:middle line:84% So the verb phrase immediately dominates the V and NP. 01:05:18.500 --> 01:05:22.730 align:middle line:84% That's just another scarier way to say that the VP is 01:05:22.730 --> 01:05:24.620 align:middle line:90% the mother of the V and the NP. 01:05:24.620 --> 01:05:28.880 align:middle line:84% The VP is the thing you created by merging V and NP together. 01:05:28.880 --> 01:05:33.350 align:middle line:84% And dominate-- so immediately dominate 01:05:33.350 --> 01:05:36.330 align:middle line:84% is kind of the basic relation in trees. 01:05:36.330 --> 01:05:38.750 align:middle line:84% It's the relation that's created by merge. 01:05:38.750 --> 01:05:42.260 align:middle line:84% Dominate is the transitive closure of immediate dominate. 01:05:42.260 --> 01:05:46.250 align:middle line:84% So VP dominates everything it immediately dominates, 01:05:46.250 --> 01:05:48.350 align:middle line:84% and everything those things immediately dominate, 01:05:48.350 --> 01:05:49.730 align:middle line:84% and everything those things immediately 01:05:49.730 --> 01:05:51.897 align:middle line:84% dominate all the way down to the bottom of the tree. 01:05:51.897 --> 01:05:54.680 align:middle line:84% So that verb phrase dominates the verb and the NP 01:05:54.680 --> 01:05:57.320 align:middle line:84% and also the D and the N-- "the" and "child." 01:05:57.320 --> 01:06:01.333 align:middle line:84% It dominates the things that are below it that were there 01:06:01.333 --> 01:06:02.375 align:middle line:90% when it was put together. 01:06:02.375 --> 01:06:06.140 align:middle line:90% 01:06:06.140 --> 01:06:09.140 align:middle line:84% And I've been using this word "constituent," 01:06:09.140 --> 01:06:11.120 align:middle line:84% but now we can talk about constituents 01:06:11.120 --> 01:06:12.170 align:middle line:90% using this terminology. 01:06:12.170 --> 01:06:13.970 align:middle line:84% We'll say that something is a constituent 01:06:13.970 --> 01:06:17.030 align:middle line:84% if all and only the words in that thing, 01:06:17.030 --> 01:06:19.140 align:middle line:84% alpha, are dominated by a single node. 01:06:19.140 --> 01:06:21.950 align:middle line:84% So there is a constituent-- "tickle the child," 01:06:21.950 --> 01:06:23.450 align:middle line:90% for example-- that verb phrase-- 01:06:23.450 --> 01:06:25.520 align:middle line:84% because there is a verb phrase that 01:06:25.520 --> 01:06:28.230 align:middle line:84% dominates just those words, "tickle the child." 01:06:28.230 --> 01:06:31.580 align:middle line:84% There is no constituent "will tickle" in this tree, 01:06:31.580 --> 01:06:33.470 align:middle line:84% because there's no node that just dominates 01:06:33.470 --> 01:06:34.490 align:middle line:90% the words "will tickle." 01:06:34.490 --> 01:06:40.560 align:middle line:90% 01:06:40.560 --> 01:06:42.763 align:middle line:84% Let me-- because we are just about out of time, 01:06:42.763 --> 01:06:45.180 align:middle line:84% but not quite-- and this would get us into something new-- 01:06:45.180 --> 01:06:48.150 align:middle line:84% let me get started on getting rid of this tree, 01:06:48.150 --> 01:06:50.460 align:middle line:84% and then we'll pick it up here next time. 01:06:50.460 --> 01:06:54.120 align:middle line:90% Let me show you some data. 01:06:54.120 --> 01:07:00.840 align:middle line:84% Consider sentences like "She likes Mary." 01:07:00.840 --> 01:07:03.050 align:middle line:90% Let's start with that one. 01:07:03.050 --> 01:07:03.550 align:middle line:90% No-- here. 01:07:03.550 --> 01:07:12.660 align:middle line:84% Let's do "She thinks I like Mary," and for that matter, 01:07:12.660 --> 01:07:17.010 align:middle line:90% "She thinks I"-- 01:07:17.010 --> 01:07:23.110 align:middle line:90% no-- "Mary thinks I like her." 01:07:23.110 --> 01:07:25.110 align:middle line:84% These are two grammatical sentences in English-- 01:07:25.110 --> 01:07:27.360 align:middle line:84% "She thinks I like Mary" and "Mary thinks I like her." 01:07:27.360 --> 01:07:30.120 align:middle line:90% 01:07:30.120 --> 01:07:33.428 align:middle line:84% In the second sentence, we've got "Mary" 01:07:33.428 --> 01:07:34.470 align:middle line:90% and we've also got "her." 01:07:34.470 --> 01:07:35.935 align:middle line:90% Can "her" refer to Mary? 01:07:35.935 --> 01:07:36.560 align:middle line:90% AUDIENCE: Yeah. 01:07:36.560 --> 01:07:37.477 align:middle line:90% NORVIN RICHARDS: Yeah. 01:07:37.477 --> 01:07:40.610 align:middle line:84% It could also be somebody else, but it could be Mary. 01:07:40.610 --> 01:07:43.880 align:middle line:84% We'll mark that by putting a little subscript on "Mary." 01:07:43.880 --> 01:07:45.560 align:middle line:84% I'll call it subscript I, and we'll 01:07:45.560 --> 01:07:47.960 align:middle line:90% put the same subscript on her. 01:07:47.960 --> 01:07:49.820 align:middle line:84% So "her" could refer back to "Mary." 01:07:49.820 --> 01:07:52.280 align:middle line:84% It could also refer to any other person-- 01:07:52.280 --> 01:07:55.270 align:middle line:90% any other female person. 01:07:55.270 --> 01:07:56.860 align:middle line:90% "She thinks I like Mary." 01:07:56.860 --> 01:07:59.020 align:middle line:90% Can "she" be Mary? 01:07:59.020 --> 01:08:00.460 align:middle line:90% No. 01:08:00.460 --> 01:08:03.490 align:middle line:90% So here, she-- 01:08:03.490 --> 01:08:05.602 align:middle line:90% Mary can't be the same person. 01:08:05.602 --> 01:08:07.810 align:middle line:84% It would be bad for "she" to refer to the same person 01:08:07.810 --> 01:08:09.310 align:middle line:90% as "Mary." 01:08:09.310 --> 01:08:10.600 align:middle line:90% Why is that do you think? 01:08:10.600 --> 01:08:14.340 align:middle line:90% 01:08:14.340 --> 01:08:15.610 align:middle line:90% Yeah? 01:08:15.610 --> 01:08:17.979 align:middle line:84% AUDIENCE: Because that's [INAUDIBLE] "I like Mary" 01:08:17.979 --> 01:08:23.140 align:middle line:84% is a constituent, and so when I define "Mary" inside 01:08:23.140 --> 01:08:26.479 align:middle line:84% of the constituent-- oh, shoot-- this is not working out. 01:08:26.479 --> 01:08:27.479 align:middle line:90% NORVIN RICHARDS: No, no. 01:08:27.479 --> 01:08:29.133 align:middle line:90% AUDIENCE: [INAUDIBLE]. 01:08:29.133 --> 01:08:31.300 align:middle line:84% NORVIN RICHARDS: Does anybody have a simpler theory? 01:08:31.300 --> 01:08:31.750 align:middle line:90% Yeah? 01:08:31.750 --> 01:08:33.958 align:middle line:84% AUDIENCE: Is it simply because in the first sentence, 01:08:33.958 --> 01:08:36.790 align:middle line:84% Mary hasn't been established as a person yet [INAUDIBLE]?? 01:08:36.790 --> 01:08:38.060 align:middle line:90% NORVIN RICHARDS: So good. 01:08:38.060 --> 01:08:40.227 align:middle line:84% That's where I was hoping the first person would go. 01:08:40.227 --> 01:08:43.000 align:middle line:84% Joseph is already going more sophisticated. 01:08:43.000 --> 01:08:46.029 align:middle line:84% So one theory we could have would be "Mary" 01:08:46.029 --> 01:08:49.689 align:middle line:84% can't refer to somebody-- so "she" can't refer to somebody 01:08:49.689 --> 01:08:51.340 align:middle line:90% who hasn't been mentioned yet. 01:08:51.340 --> 01:08:55.750 align:middle line:84% She can't refer to somebody who's later in the tree. 01:08:55.750 --> 01:09:01.240 align:middle line:84% Here, "her" is referring to somebody on its left. 01:09:01.240 --> 01:09:09.430 align:middle line:84% Notice-- think about a sentence like "Everyone 01:09:09.430 --> 01:09:17.664 align:middle line:90% who knows her likes Mary." 01:09:17.664 --> 01:09:20.680 align:middle line:90% Can her refer to Mary there? 01:09:20.680 --> 01:09:21.310 align:middle line:90% Yeah. 01:09:21.310 --> 01:09:24.399 align:middle line:84% So not to immediately dump on your hypothesis, 01:09:24.399 --> 01:09:26.810 align:middle line:84% but I'm immediately dumping on your hypothesis. 01:09:26.810 --> 01:09:29.770 align:middle line:84% So pronouns actually can refer to things 01:09:29.770 --> 01:09:32.740 align:middle line:90% that are later in the tree. 01:09:32.740 --> 01:09:35.290 align:middle line:84% Sorry, let me say that again more coherently. 01:09:35.290 --> 01:09:38.050 align:middle line:84% Pronouns can refer to things that are later in the sentence. 01:09:38.050 --> 01:09:40.258 align:middle line:84% You can tell I'm a syntactician, because when I think 01:09:40.258 --> 01:09:42.790 align:middle line:90% of sentences, I think of trees. 01:09:42.790 --> 01:09:49.810 align:middle line:84% So the relationship that holds between "she" and "Mary" here-- 01:09:49.810 --> 01:09:53.050 align:middle line:90% between "her" and "Mary" here-- 01:09:53.050 --> 01:09:56.350 align:middle line:84% the thing that tells you whether a pronoun can 01:09:56.350 --> 01:09:58.040 align:middle line:90% refer to a person-- 01:09:58.040 --> 01:10:01.360 align:middle line:84% it can't just be about whether the person's name is later 01:10:01.360 --> 01:10:02.440 align:middle line:90% in the sentence. 01:10:02.440 --> 01:10:04.720 align:middle line:84% It's going to turn out to be about the structure 01:10:04.720 --> 01:10:05.890 align:middle line:90% of the sentence. 01:10:05.890 --> 01:10:08.140 align:middle line:84% So we'll develop a theory of these kinds 01:10:08.140 --> 01:10:10.630 align:middle line:84% of facts, which will be another test 01:10:10.630 --> 01:10:12.580 align:middle line:90% for constituent structures. 01:10:12.580 --> 01:10:15.940 align:middle line:84% We'll develop a theory of that, and it will tell us 01:10:15.940 --> 01:10:19.420 align:middle line:90% that this tree is wrong. 01:10:19.420 --> 01:10:23.620 align:middle line:84% But I'll first finish developing that theory, 01:10:23.620 --> 01:10:27.040 align:middle line:84% and then we'll be able to use it as an additional test 01:10:27.040 --> 01:10:29.710 align:middle line:90% for structure. 01:10:29.710 --> 01:10:32.500 align:middle line:84% That's the incentive to come to class next time. 01:10:32.500 --> 01:10:35.290 align:middle line:84% You'll get to see how that's going to work. 01:10:35.290 --> 01:10:37.060 align:middle line:84% Let's stop there unless people have 01:10:37.060 --> 01:10:39.140 align:middle line:90% any questions about this stuff. 01:10:39.140 --> 01:10:39.640 align:middle line:90% Good. 01:10:39.640 --> 01:10:41.290 align:middle line:84% Let's just pause it right there, because if I 01:10:41.290 --> 01:10:43.040 align:middle line:84% start trying to show you this any further, 01:10:43.040 --> 01:10:45.240 align:middle line:90% we'll just all get confused. 01:10:45.240 --> 01:10:49.000 align:middle line:90%