WEBVTT 00:00:00.000 --> 00:00:01.960 align:middle line:90% [SQUEAKING] 00:00:01.960 --> 00:00:03.430 align:middle line:90% [RUSTLING] 00:00:03.430 --> 00:00:06.370 align:middle line:90% [CLICKING] 00:00:06.370 --> 00:00:11.760 align:middle line:90% 00:00:11.760 --> 00:00:12.760 align:middle line:90% NORVIN RICHARDS: OK. 00:00:12.760 --> 00:00:17.920 align:middle line:84% So last time, we ended on one of these cliffhangers. 00:00:17.920 --> 00:00:21.520 align:middle line:84% And so I rushed you through a bunch of Japanese data. 00:00:21.520 --> 00:00:26.185 align:middle line:84% And then we got to this new bit of linguistic technology-- 00:00:26.185 --> 00:00:28.720 align:middle line:84% [BUZZING] that's going to be annoying-- 00:00:28.720 --> 00:00:30.400 align:middle line:84% this new bit of linguistic technology 00:00:30.400 --> 00:00:33.310 align:middle line:90% and went through it pretty fast. 00:00:33.310 --> 00:00:37.510 align:middle line:84% So what I want to do is do the run up to that again a little 00:00:37.510 --> 00:00:40.888 align:middle line:84% more carefully and show you what it is that we're talking about. 00:00:40.888 --> 00:00:42.430 align:middle line:84% What we were talking about at the end 00:00:42.430 --> 00:00:46.430 align:middle line:84% was the position of accent in Japanese compounds. 00:00:46.430 --> 00:00:50.690 align:middle line:84% So here are a bunch of compound nouns in Japanese. 00:00:50.690 --> 00:00:52.930 align:middle line:84% And the observation was that when 00:00:52.930 --> 00:00:58.780 align:middle line:84% you have a compound consisting of two parts, 00:00:58.780 --> 00:01:00.700 align:middle line:84% the result always has one accent. 00:01:00.700 --> 00:01:03.790 align:middle line:84% It has accent on one particular syllable. 00:01:03.790 --> 00:01:06.040 align:middle line:84% The accented syllable doesn't have 00:01:06.040 --> 00:01:10.480 align:middle line:84% to be the accented syllable in either of the original parts 00:01:10.480 --> 00:01:11.360 align:middle line:90% of the compound. 00:01:11.360 --> 00:01:15.130 align:middle line:84% So "raw egg," which is the first example in this slide-- 00:01:15.130 --> 00:01:19.020 align:middle line:84% "raw" and "egg"-- they have accents 00:01:19.020 --> 00:01:20.020 align:middle line:90% where they have accents. 00:01:20.020 --> 00:01:22.180 align:middle line:84% "Raw" has an accent on the first syllable. 00:01:22.180 --> 00:01:24.100 align:middle line:84% "Egg" has an accent on the second syllable. 00:01:24.100 --> 00:01:27.100 align:middle line:84% But "raw egg" has an accent on the syllable which is not 00:01:27.100 --> 00:01:29.830 align:middle line:90% normally accented in "egg." 00:01:29.830 --> 00:01:33.940 align:middle line:84% What we said was there is a set of rules for where 00:01:33.940 --> 00:01:36.430 align:middle line:90% accent goes in these compounds. 00:01:36.430 --> 00:01:39.310 align:middle line:84% And the rules go something like this. 00:01:39.310 --> 00:01:41.380 align:middle line:84% There's only going to be one accent. 00:01:41.380 --> 00:01:45.370 align:middle line:84% It needs to be next to the boundary between the words. 00:01:45.370 --> 00:01:47.860 align:middle line:84% And if possible, it should be on the second word. 00:01:47.860 --> 00:01:51.040 align:middle line:84% That is, it should be after the boundary between the words. 00:01:51.040 --> 00:01:54.100 align:middle line:84% But you don't do that if that would result in the accent 00:01:54.100 --> 00:01:55.450 align:middle line:90% being on the last syllable. 00:01:55.450 --> 00:01:59.110 align:middle line:84% So the general pattern is the one you're getting in "French 00:01:59.110 --> 00:02:01.540 align:middle line:84% fry" up there-- "furaido POteto"-- 00:02:01.540 --> 00:02:04.960 align:middle line:84% where you get accent at the beginning of the second half 00:02:04.960 --> 00:02:06.310 align:middle line:90% of the compound, "potato." 00:02:06.310 --> 00:02:07.930 align:middle line:90% Yeah. 00:02:07.930 --> 00:02:10.539 align:middle line:84% That's what you get in most of the examples on the preceding 00:02:10.539 --> 00:02:11.039 align:middle line:90% slide. 00:02:11.039 --> 00:02:12.890 align:middle line:84% If we go back to the preceding slide, 00:02:12.890 --> 00:02:14.498 align:middle line:84% you can see that in "raw egg," right? 00:02:14.498 --> 00:02:15.790 align:middle line:90% It's at the beginning of "egg." 00:02:15.790 --> 00:02:17.020 align:middle line:84% And you can see it in "field mouse." 00:02:17.020 --> 00:02:18.395 align:middle line:84% It's at the beginning of "mouse." 00:02:18.395 --> 00:02:21.340 align:middle line:84% And you're also getting it in "French fry." 00:02:21.340 --> 00:02:25.180 align:middle line:84% But there's this coherent class of exceptions. 00:02:25.180 --> 00:02:27.610 align:middle line:84% If the second half of the compound 00:02:27.610 --> 00:02:29.830 align:middle line:84% is only one syllable long, then you 00:02:29.830 --> 00:02:31.390 align:middle line:84% don't put accent on the second half. 00:02:31.390 --> 00:02:33.920 align:middle line:84% You put it at the end of the first half. 00:02:33.920 --> 00:02:38.710 align:middle line:84% So Kagawa prefecture-- "kagawa" is the name of a place. 00:02:38.710 --> 00:02:41.410 align:middle line:84% And it's got accent on its first syllable. 00:02:41.410 --> 00:02:44.560 align:middle line:84% But the expression for "Kagawa prefecture" 00:02:44.560 --> 00:02:47.467 align:middle line:84% has accent at almost at the boundary. 00:02:47.467 --> 00:02:49.300 align:middle line:84% It's right before the boundary between them, 00:02:49.300 --> 00:02:52.025 align:middle line:84% not on the second half of the compound, unusually, 00:02:52.025 --> 00:02:53.650 align:middle line:84% because the second half of the compound 00:02:53.650 --> 00:02:54.760 align:middle line:90% is only one syllable long. 00:02:54.760 --> 00:02:56.853 align:middle line:90% So we get "kagaWAken." 00:02:56.853 --> 00:02:59.220 align:middle line:90% Yeah. 00:02:59.220 --> 00:03:02.010 align:middle line:84% I'm telling you all this not just to educate you 00:03:02.010 --> 00:03:05.430 align:middle line:84% in what a cool and complicated language Japanese is, 00:03:05.430 --> 00:03:07.560 align:middle line:84% but to show you that there are cases 00:03:07.560 --> 00:03:09.930 align:middle line:84% of cool and complicated phenomena 00:03:09.930 --> 00:03:13.440 align:middle line:84% that can be fruitfully thought of as a series 00:03:13.440 --> 00:03:16.380 align:middle line:84% of things that you're trying to do which are in conflict 00:03:16.380 --> 00:03:17.880 align:middle line:90% with each other sometimes. 00:03:17.880 --> 00:03:22.740 align:middle line:84% So Japanese wants the accent to be on the second word 00:03:22.740 --> 00:03:24.810 align:middle line:84% and next to the boundary between the words. 00:03:24.810 --> 00:03:27.690 align:middle line:84% But it also wants it not to be final. 00:03:27.690 --> 00:03:31.650 align:middle line:84% And if it can't have all of those things at once, 00:03:31.650 --> 00:03:33.810 align:middle line:84% like in "Kagawa prefecture," it has 00:03:33.810 --> 00:03:36.780 align:middle line:84% to decide, do I care more about it not being final, 00:03:36.780 --> 00:03:38.970 align:middle line:84% or do I care more about it being on the second word? 00:03:38.970 --> 00:03:42.930 align:middle line:84% And Japanese cares more about it not being final, right? 00:03:42.930 --> 00:03:44.640 align:middle line:84% So it retracts it from the second word 00:03:44.640 --> 00:03:47.468 align:middle line:90% to the end of the first word. 00:03:47.468 --> 00:03:48.510 align:middle line:90% Does anybody still read-- 00:03:48.510 --> 00:03:51.530 align:middle line:84% Isaac Asimov had these stories, the robot stories. 00:03:51.530 --> 00:03:53.280 align:middle line:90% He had a book, I, Robot. 00:03:53.280 --> 00:03:54.722 align:middle line:90% It was a series of stories. 00:03:54.722 --> 00:03:56.430 align:middle line:84% Some of you have heard of them, at least? 00:03:56.430 --> 00:03:56.930 align:middle line:90% Yeah. 00:03:56.930 --> 00:04:01.830 align:middle line:84% So they feature robots that have these laws governing 00:04:01.830 --> 00:04:02.760 align:middle line:90% how they work. 00:04:02.760 --> 00:04:04.980 align:middle line:84% And the laws act like these kinds of laws. 00:04:04.980 --> 00:04:06.870 align:middle line:90% So there's the first law. 00:04:06.870 --> 00:04:08.340 align:middle line:90% And then there's the second law. 00:04:08.340 --> 00:04:09.757 align:middle line:84% And the rule for the second law is 00:04:09.757 --> 00:04:11.638 align:middle line:84% you must obey the second law unless it comes 00:04:11.638 --> 00:04:12.930 align:middle line:90% in conflict with the first law. 00:04:12.930 --> 00:04:15.090 align:middle line:84% Then you disobey the second law and do the first law, right? 00:04:15.090 --> 00:04:16.707 align:middle line:84% And there's a third law, which you 00:04:16.707 --> 00:04:18.540 align:middle line:84% should do unless it conflicts with the first 00:04:18.540 --> 00:04:19.620 align:middle line:90% or the second law. 00:04:19.620 --> 00:04:21.390 align:middle line:90% So this is like that, right? 00:04:21.390 --> 00:04:26.760 align:middle line:84% So Japanese wants the accent to be on the second word, oh, 00:04:26.760 --> 00:04:28.770 align:middle line:84% unless that's going to make the accent final. 00:04:28.770 --> 00:04:30.840 align:middle line:90% Then you don't do that. 00:04:30.840 --> 00:04:33.480 align:middle line:90% Does that make sense? 00:04:33.480 --> 00:04:37.530 align:middle line:84% That's a way-- or to put that another way, don't 00:04:37.530 --> 00:04:39.150 align:middle line:84% put the accent on the last syllable 00:04:39.150 --> 00:04:42.355 align:middle line:84% is more important than put the accent on the second word. 00:04:42.355 --> 00:04:43.980 align:middle line:84% That's the picture that we end up with. 00:04:43.980 --> 00:04:50.820 align:middle line:84% So you get "kagaWAken" with the accent at the end of "kagawa" 00:04:50.820 --> 00:04:53.280 align:middle line:90% and not "kagawaKEN." 00:04:53.280 --> 00:04:57.810 align:middle line:90% You don't put accent at the end. 00:04:57.810 --> 00:04:59.730 align:middle line:84% There's a standard way of representing 00:04:59.730 --> 00:05:02.400 align:middle line:84% these kinds of facts, these bodies of facts, which 00:05:02.400 --> 00:05:04.650 align:middle line:84% involves this table that I've got here 00:05:04.650 --> 00:05:07.080 align:middle line:90% at the bottom of the slide. 00:05:07.080 --> 00:05:09.930 align:middle line:84% This was an approach to phonology that was 00:05:09.930 --> 00:05:13.860 align:middle line:90% invented by people who-- 00:05:13.860 --> 00:05:15.960 align:middle line:84% one of them in particular, Alan Prince, 00:05:15.960 --> 00:05:18.360 align:middle line:90% is a brilliant, brilliant man. 00:05:18.360 --> 00:05:20.380 align:middle line:84% But he enjoys intimidating people. 00:05:20.380 --> 00:05:23.100 align:middle line:84% And so instead of calling this a table, 00:05:23.100 --> 00:05:25.380 align:middle line:90% he decided to call it a tableau. 00:05:25.380 --> 00:05:28.410 align:middle line:84% So he gives it a French name in order 00:05:28.410 --> 00:05:31.050 align:middle line:84% to be as intimidating as possible. 00:05:31.050 --> 00:05:35.940 align:middle line:84% So I'll adopt that because that's what phonologists, 00:05:35.940 --> 00:05:38.620 align:middle line:84% linguists standardly call these objects. 00:05:38.620 --> 00:05:40.470 align:middle line:90% This is a tableau. 00:05:40.470 --> 00:05:43.680 align:middle line:84% And what the tableau represents is this interplay 00:05:43.680 --> 00:05:46.350 align:middle line:84% between the different laws that are governing 00:05:46.350 --> 00:05:48.510 align:middle line:90% where the accent goes. 00:05:48.510 --> 00:05:52.560 align:middle line:84% The idea is the most important laws are toward the left. 00:05:52.560 --> 00:05:56.130 align:middle line:84% And the least important laws are toward the right. 00:05:56.130 --> 00:05:58.950 align:middle line:84% And what you do is imagine all of the places 00:05:58.950 --> 00:06:00.420 align:middle line:90% where you could put the accent. 00:06:00.420 --> 00:06:05.160 align:middle line:84% And you ask yourself, how does that placement of the accent 00:06:05.160 --> 00:06:08.040 align:middle line:84% square with the various different laws? 00:06:08.040 --> 00:06:12.750 align:middle line:84% The little asterisks mark places where one of the candidates, 00:06:12.750 --> 00:06:15.870 align:middle line:84% one of the options you're considering, 00:06:15.870 --> 00:06:18.040 align:middle line:90% violates one of the laws. 00:06:18.040 --> 00:06:23.130 align:middle line:84% So for example, the first one, which kept both of the accents 00:06:23.130 --> 00:06:27.420 align:middle line:84% in the parts of the compound, just kept them both, 00:06:27.420 --> 00:06:29.160 align:middle line:84% violates the most important law, which is 00:06:29.160 --> 00:06:30.540 align:middle line:90% you may only have one accent. 00:06:30.540 --> 00:06:32.920 align:middle line:90% So it gets a star right there. 00:06:32.920 --> 00:06:35.040 align:middle line:84% And that eliminates that one from competition 00:06:35.040 --> 00:06:38.700 align:middle line:84% because it's violated the most important law, the one that's 00:06:38.700 --> 00:06:40.840 align:middle line:90% furthest on the left. 00:06:40.840 --> 00:06:44.400 align:middle line:84% And then what I've done is consider examples 00:06:44.400 --> 00:06:46.830 align:middle line:84% where the accent is-- it's a four-syllable word, so we're 00:06:46.830 --> 00:06:50.040 align:middle line:84% considering putting the single accent on any of the four 00:06:50.040 --> 00:06:51.000 align:middle line:90% syllables. 00:06:51.000 --> 00:06:54.600 align:middle line:84% The first two, which have the accent 00:06:54.600 --> 00:06:56.100 align:middle line:84% on the first or the second syllable, 00:06:56.100 --> 00:06:57.530 align:middle line:84% violate the second law, which says 00:06:57.530 --> 00:06:59.530 align:middle line:84% put the accent near the boundary between the two 00:06:59.530 --> 00:07:00.730 align:middle line:90% halves of the compound. 00:07:00.730 --> 00:07:04.350 align:middle line:84% So those are ones in which the accent is not 00:07:04.350 --> 00:07:07.630 align:middle line:84% on a syllable that's next to the boundary. 00:07:07.630 --> 00:07:13.140 align:middle line:84% And then the next to last one, which is the one that wins-- 00:07:13.140 --> 00:07:15.632 align:middle line:84% that's the one that obeys all of the conditions 00:07:15.632 --> 00:07:17.340 align:middle line:84% except for the least important one, which 00:07:17.340 --> 00:07:21.300 align:middle line:84% is it's best to put your accent on the last word, right? 00:07:21.300 --> 00:07:23.910 align:middle line:90% 00:07:23.910 --> 00:07:26.220 align:middle line:84% You notice, there is no perfect candidate. 00:07:26.220 --> 00:07:29.040 align:middle line:84% There isn't anything that doesn't have any stars. 00:07:29.040 --> 00:07:30.630 align:middle line:84% What the tableau does is allow us 00:07:30.630 --> 00:07:35.730 align:middle line:84% to think in a formal way about how 00:07:35.730 --> 00:07:39.218 align:middle line:84% you satisfy conditions that are not compatible with each other. 00:07:39.218 --> 00:07:41.010 align:middle line:84% And I just realized I get to take this off. 00:07:41.010 --> 00:07:43.680 align:middle line:90% I keep forgetting that. 00:07:43.680 --> 00:07:45.100 align:middle line:90% Yeah. 00:07:45.100 --> 00:07:47.770 align:middle line:84% So that's what the tableau is meant to represent. 00:07:47.770 --> 00:07:50.080 align:middle line:84% Are there are questions about the tableau? 00:07:50.080 --> 00:07:52.750 align:middle line:84% There's a standard way to represent the fact 00:07:52.750 --> 00:07:55.337 align:middle line:84% that the winner is that next to last one 00:07:55.337 --> 00:07:57.670 align:middle line:84% there, where the stress is on the next to last syllable, 00:07:57.670 --> 00:07:59.462 align:middle line:84% the accent is on the next to last syllable, 00:07:59.462 --> 00:08:02.240 align:middle line:84% which is to put a pointing finger next to it like that. 00:08:02.240 --> 00:08:04.780 align:middle line:90% This is the one that wins. 00:08:04.780 --> 00:08:08.980 align:middle line:84% Optimality theory-- this approach to phonology 00:08:08.980 --> 00:08:10.520 align:middle line:90% is called optimality theory. 00:08:10.520 --> 00:08:12.430 align:middle line:90% And I don't know. 00:08:12.430 --> 00:08:14.080 align:middle line:84% I guess they just liked the font that 00:08:14.080 --> 00:08:15.330 align:middle line:90% had the pointing finger in it. 00:08:15.330 --> 00:08:17.594 align:middle line:90% So they decided to use it. 00:08:17.594 --> 00:08:18.137 align:middle line:90% OK? 00:08:18.137 --> 00:08:18.970 align:middle line:90% This is an approach. 00:08:18.970 --> 00:08:19.906 align:middle line:90% Yeah? 00:08:19.906 --> 00:08:24.710 align:middle line:84% STUDENT: Question about-- there is an accent on-- 00:08:24.710 --> 00:08:27.570 align:middle line:84% how do we know there's an accent on "ken" 00:08:27.570 --> 00:08:30.593 align:middle line:84% if it just never shows up because every time [INAUDIBLE]?? 00:08:30.593 --> 00:08:32.510 align:middle line:84% NORVIN RICHARDS: Oh, you mean why am I putting 00:08:32.510 --> 00:08:34.429 align:middle line:90% an accent on "ken" over there? 00:08:34.429 --> 00:08:36.770 align:middle line:84% STUDENT: Because every time "ken" shows up, 00:08:36.770 --> 00:08:39.000 align:middle line:90% one syllable is at the end. 00:08:39.000 --> 00:08:42.130 align:middle line:84% So there's no-- so it's always the final accent. 00:08:42.130 --> 00:08:43.630 align:middle line:84% So how do we know there's an accent? 00:08:43.630 --> 00:08:46.190 align:middle line:84% NORVIN RICHARDS: Yeah, that's a very good point. 00:08:46.190 --> 00:08:48.790 align:middle line:84% It is possible to pronounce "ken" by itself, 00:08:48.790 --> 00:08:50.720 align:middle line:90% not as part of a compound. 00:08:50.720 --> 00:08:53.590 align:middle line:84% So if you want to talk about "the prefecture," it's a word. 00:08:53.590 --> 00:08:56.770 align:middle line:84% It means "prefecture," which is like a state. 00:08:56.770 --> 00:09:00.310 align:middle line:84% It's an administrative subdivision of Japan. 00:09:00.310 --> 00:09:02.303 align:middle line:84% And it's possible to say that word by itself. 00:09:02.303 --> 00:09:03.220 align:middle line:90% STUDENT: So unbounded. 00:09:03.220 --> 00:09:04.178 align:middle line:90% NORVIN RICHARDS: Sorry? 00:09:04.178 --> 00:09:05.850 align:middle line:90% STUDENT: So it's unfounded? 00:09:05.850 --> 00:09:08.020 align:middle line:90% NORVIN RICHARDS: It's? 00:09:08.020 --> 00:09:09.020 align:middle line:90% STUDENT: It's unbounded? 00:09:09.020 --> 00:09:11.300 align:middle line:84% NORVIN RICHARDS: "ken" is unbounded? 00:09:11.300 --> 00:09:13.530 align:middle line:84% I don't think I know what you mean. 00:09:13.530 --> 00:09:14.780 align:middle line:90% Oh, it's not a bound morpheme. 00:09:14.780 --> 00:09:15.080 align:middle line:90% That's right. 00:09:15.080 --> 00:09:16.220 align:middle line:90% It can be a free morpheme. 00:09:16.220 --> 00:09:16.938 align:middle line:90% Yes, yes. 00:09:16.938 --> 00:09:17.480 align:middle line:90% That's right. 00:09:17.480 --> 00:09:21.500 align:middle line:84% So you can say in Japanese, "The prefecture is large" 00:09:21.500 --> 00:09:22.740 align:middle line:90% or whatever. 00:09:22.740 --> 00:09:27.160 align:middle line:84% And then you can show that it's accented, yeah. 00:09:27.160 --> 00:09:28.100 align:middle line:90% Good question. 00:09:28.100 --> 00:09:31.700 align:middle line:84% There are other questions people have about this? 00:09:31.700 --> 00:09:32.690 align:middle line:90% OK. 00:09:32.690 --> 00:09:36.170 align:middle line:84% So this is just meant to illustrate the workings 00:09:36.170 --> 00:09:38.840 align:middle line:90% of optimality theory. 00:09:38.840 --> 00:09:40.870 align:middle line:90% Why am I showing you this one? 00:09:40.870 --> 00:09:41.780 align:middle line:90% I'm not sure. 00:09:41.780 --> 00:09:43.050 align:middle line:90% OK. 00:09:43.050 --> 00:09:44.010 align:middle line:90% Oh, I see. 00:09:44.010 --> 00:09:47.490 align:middle line:84% This is a similar tableau for "French fry." 00:09:47.490 --> 00:09:51.480 align:middle line:84% So this has the same constraints, 00:09:51.480 --> 00:09:55.530 align:middle line:84% these things that I put in all caps up at the top. 00:09:55.530 --> 00:09:59.070 align:middle line:84% In this case, you do end up with accent 00:09:59.070 --> 00:10:01.650 align:middle line:84% at the beginning of potato on the second word. 00:10:01.650 --> 00:10:03.610 align:middle line:84% That actually is a perfect candidate. 00:10:03.610 --> 00:10:06.990 align:middle line:84% So this is an example in which none of the laws 00:10:06.990 --> 00:10:09.757 align:middle line:84% are violated by the best candidate. 00:10:09.757 --> 00:10:11.590 align:middle line:84% The laws aren't in conflict with each other. 00:10:11.590 --> 00:10:13.173 align:middle line:84% So you can put accent at the beginning 00:10:13.173 --> 00:10:15.450 align:middle line:90% of the last word, "potato." 00:10:15.450 --> 00:10:17.160 align:middle line:90% And it won't be final. 00:10:17.160 --> 00:10:18.930 align:middle line:90% Yeah, yeah? 00:10:18.930 --> 00:10:22.224 align:middle line:84% STUDENT: So for example, if this perfect candidate didn't exist, 00:10:22.224 --> 00:10:24.182 align:middle line:84% and you were having to choose between the first 00:10:24.182 --> 00:10:24.930 align:middle line:90% and the third-- 00:10:24.930 --> 00:10:25.290 align:middle line:90% NORVIN RICHARDS: Yeah. 00:10:25.290 --> 00:10:26.748 align:middle line:84% STUDENT: --you would want to choose 00:10:26.748 --> 00:10:29.392 align:middle line:84% the third because it violates a rule that's less important. 00:10:29.392 --> 00:10:31.350 align:middle line:84% NORVIN RICHARDS: Yeah, that's absolutely right. 00:10:31.350 --> 00:10:33.870 align:middle line:84% So if, for some reason, that fourth candidate 00:10:33.870 --> 00:10:36.390 align:middle line:84% were impossible-- say, there was another very high 00:10:36.390 --> 00:10:43.980 align:middle line:84% ranking constraint-- don't put accent on "potato," then yes. 00:10:43.980 --> 00:10:45.240 align:middle line:90% The third would win. 00:10:45.240 --> 00:10:45.990 align:middle line:90% That's right. 00:10:45.990 --> 00:10:48.508 align:middle line:84% And actually, we can see that by looking 00:10:48.508 --> 00:10:49.800 align:middle line:90% at the behavior of "kagaWAken." 00:10:49.800 --> 00:10:53.880 align:middle line:84% So if we eliminate the option of putting accent 00:10:53.880 --> 00:10:56.400 align:middle line:84% on the last word, that's where the accent goes. 00:10:56.400 --> 00:10:58.630 align:middle line:84% It goes at the end of the first word. 00:10:58.630 --> 00:11:01.950 align:middle line:84% So it's the best you can do, right? 00:11:01.950 --> 00:11:05.957 align:middle line:84% Still near the boundary, still only one accent, just 00:11:05.957 --> 00:11:07.540 align:middle line:84% didn't manage to accent the last word, 00:11:07.540 --> 00:11:08.832 align:middle line:90% which is what you'd like to do. 00:11:08.832 --> 00:11:11.880 align:middle line:84% But it's not high on your list of constraints. 00:11:11.880 --> 00:11:12.735 align:middle line:90% Yep. 00:11:12.735 --> 00:11:14.410 align:middle line:90% Does that make sense? 00:11:14.410 --> 00:11:18.510 align:middle line:84% So this is an approach to thinking about phonology. 00:11:18.510 --> 00:11:21.190 align:middle line:84% Optimality theory is the name for it. 00:11:21.190 --> 00:11:24.090 align:middle line:84% It says-- in a way, it's a comforting theory-- 00:11:24.090 --> 00:11:25.690 align:middle line:90% It says, "Do the best you can." 00:11:25.690 --> 00:11:27.108 align:middle line:90% You have all of these things. 00:11:27.108 --> 00:11:28.650 align:middle line:84% Probably, all of you have encountered 00:11:28.650 --> 00:11:30.330 align:middle line:84% things like this in other domains 00:11:30.330 --> 00:11:32.580 align:middle line:84% at MIT if you've ever dealt with engineering problems. 00:11:32.580 --> 00:11:35.588 align:middle line:84% There are probably cases where you have various things 00:11:35.588 --> 00:11:36.630 align:middle line:90% that you want to achieve. 00:11:36.630 --> 00:11:39.090 align:middle line:84% And you can't optimize all of them. 00:11:39.090 --> 00:11:42.130 align:middle line:84% You have to decide which ones are the most important. 00:11:42.130 --> 00:11:44.460 align:middle line:84% And the idea here is that language is like that. 00:11:44.460 --> 00:11:47.010 align:middle line:84% There are cases where you have to decide 00:11:47.010 --> 00:11:50.850 align:middle line:84% among your various desiderata which 00:11:50.850 --> 00:11:52.950 align:middle line:90% ones you care about the most. 00:11:52.950 --> 00:11:55.170 align:middle line:84% The hope in optimality theory is that we'll 00:11:55.170 --> 00:12:00.150 align:middle line:84% end up with a list of these constraints, these laws, 00:12:00.150 --> 00:12:02.920 align:middle line:84% these rules that will be universal. 00:12:02.920 --> 00:12:05.790 align:middle line:84% And the differences between languages 00:12:05.790 --> 00:12:07.860 align:middle line:84% are just a matter of deciding what 00:12:07.860 --> 00:12:10.410 align:middle line:84% order the different constraints are in. 00:12:10.410 --> 00:12:16.770 align:middle line:84% So there could be languages in which the one accent 00:12:16.770 --> 00:12:21.840 align:middle line:84% rule is very low ranked and the first candidate maybe wins, 00:12:21.840 --> 00:12:22.340 align:middle line:90% right? 00:12:22.340 --> 00:12:24.630 align:middle line:84% You could imagine a language like that. 00:12:24.630 --> 00:12:26.800 align:middle line:84% Tokyo Japanese doesn't work that way. 00:12:26.800 --> 00:12:29.040 align:middle line:84% So differences between languages are all a matter 00:12:29.040 --> 00:12:30.900 align:middle line:90% of reranking the constraints. 00:12:30.900 --> 00:12:34.170 align:middle line:84% That's the mission statement of optimality theory, 00:12:34.170 --> 00:12:36.970 align:middle line:84% is to try to figure out how the world could work that way. 00:12:36.970 --> 00:12:37.470 align:middle line:90% Yeah? 00:12:37.470 --> 00:12:39.905 align:middle line:90% 00:12:39.905 --> 00:12:40.405 align:middle line:90% OK. 00:12:40.405 --> 00:12:43.130 align:middle line:90% 00:12:43.130 --> 00:12:45.080 align:middle line:84% Optimality theory was partly inspired 00:12:45.080 --> 00:12:48.013 align:middle line:84% by the discovery of these kinds of conspiracies 00:12:48.013 --> 00:12:50.180 align:middle line:84% that we started talking about when I was showing you 00:12:50.180 --> 00:12:51.500 align:middle line:90% some stuff about Yawelmani. 00:12:51.500 --> 00:12:54.560 align:middle line:84% And now we've seen another example from Japanese. 00:12:54.560 --> 00:12:57.320 align:middle line:84% For Yawelmani, the conspiracy was, 00:12:57.320 --> 00:12:59.540 align:middle line:84% I don't want three consonants in a row. 00:12:59.540 --> 00:13:02.210 align:middle line:84% And I'll do various things to achieve that. 00:13:02.210 --> 00:13:05.180 align:middle line:84% We said you could state rules governing the various things. 00:13:05.180 --> 00:13:07.410 align:middle line:84% And maybe in a sense, you should. 00:13:07.410 --> 00:13:10.483 align:middle line:84% But you-- the idea is, let's try to keep sight of the fact 00:13:10.483 --> 00:13:11.900 align:middle line:84% that all of these rules are trying 00:13:11.900 --> 00:13:13.520 align:middle line:90% to achieve the same thing. 00:13:13.520 --> 00:13:16.190 align:middle line:84% We want to also capture that kind of a fact. 00:13:16.190 --> 00:13:19.220 align:middle line:84% And there's something similar going on here in Japanese. 00:13:19.220 --> 00:13:22.310 align:middle line:84% One of the rules, the highest ranking rule, 00:13:22.310 --> 00:13:26.750 align:middle line:84% was "Don't have more than one accent in a word." 00:13:26.750 --> 00:13:31.190 align:middle line:84% And what we saw last time was that there are various things 00:13:31.190 --> 00:13:33.860 align:middle line:84% that Japanese does to achieve this 00:13:33.860 --> 00:13:36.150 align:middle line:90% depending on the situation. 00:13:36.150 --> 00:13:41.540 align:middle line:84% So there are words like "gurai," which 00:13:41.540 --> 00:13:45.530 align:middle line:84% I think I translated last time as even, which are accented. 00:13:45.530 --> 00:13:47.540 align:middle line:84% And if you attach them to another word that's 00:13:47.540 --> 00:13:50.373 align:middle line:84% also accented, you delete the accent on the first word, 00:13:50.373 --> 00:13:52.040 align:middle line:84% in the word you're attaching "gurai" to. 00:13:52.040 --> 00:13:54.755 align:middle line:84% And there are other words, like "made," which means "to," 00:13:54.755 --> 00:13:58.850 align:middle line:84% or "up to," "as far as," which are accented but only 00:13:58.850 --> 00:14:02.720 align:middle line:84% if they are attached to unaccented words. 00:14:02.720 --> 00:14:05.810 align:middle line:84% And then we've done all this talking about compounds. 00:14:05.810 --> 00:14:08.360 align:middle line:84% These are all different kinds of examples in which 00:14:08.360 --> 00:14:09.380 align:middle line:90% different things happen. 00:14:09.380 --> 00:14:10.963 align:middle line:84% There are different strategies, right? 00:14:10.963 --> 00:14:13.880 align:middle line:84% So for "gurai," it's "Delete the accent on the word 00:14:13.880 --> 00:14:15.680 align:middle line:90% you've attached 'gurai' to." 00:14:15.680 --> 00:14:17.570 align:middle line:84% For "made," it's "Delete the accent on 00:14:17.570 --> 00:14:19.070 align:middle line:90% 'made' for the compounds." 00:14:19.070 --> 00:14:21.233 align:middle line:84% We've now talked about this computation 00:14:21.233 --> 00:14:23.150 align:middle line:84% that you go through to try to figure out where 00:14:23.150 --> 00:14:24.710 align:middle line:90% the single accent should go. 00:14:24.710 --> 00:14:26.300 align:middle line:84% But what all of these things are doing 00:14:26.300 --> 00:14:29.360 align:middle line:84% is trying to arrange for there to just be a single accent. 00:14:29.360 --> 00:14:31.220 align:middle line:84% And one of the virtues of optimality theory 00:14:31.220 --> 00:14:36.290 align:middle line:84% is that it allows us to make that part of our analysis. 00:14:36.290 --> 00:14:38.930 align:middle line:84% If we just stated these rules, we'd 00:14:38.930 --> 00:14:41.210 align:middle line:84% be losing sight of the fact that these rules all 00:14:41.210 --> 00:14:43.468 align:middle line:84% seem to be trying to achieve the same thing. 00:14:43.468 --> 00:14:46.010 align:middle line:84% That's one of the things people who work on optimality theory 00:14:46.010 --> 00:14:46.940 align:middle line:90% are trying to get. 00:14:46.940 --> 00:14:48.260 align:middle line:84% And then of course, the goal is to try 00:14:48.260 --> 00:14:49.850 align:middle line:84% to understand why we have the constraints 00:14:49.850 --> 00:14:51.110 align:middle line:90% that we have and not others. 00:14:51.110 --> 00:14:53.690 align:middle line:84% That's another thing people who work on this try to do. 00:14:53.690 --> 00:14:59.290 align:middle line:90% 00:14:59.290 --> 00:14:59.790 align:middle line:90% Yes. 00:14:59.790 --> 00:15:02.290 align:middle line:90% I just said that and that. 00:15:02.290 --> 00:15:05.790 align:middle line:84% So remember, Yawelmani, which was "Avoid strings of three 00:15:05.790 --> 00:15:09.210 align:middle line:84% consonants either by inserting vowels or by deleting 00:15:09.210 --> 00:15:12.330 align:middle line:90% 'h' if you have an 'h'." 00:15:12.330 --> 00:15:18.360 align:middle line:84% So I've been saying we could just state these rules. 00:15:18.360 --> 00:15:20.340 align:middle line:84% But we'd be missing a generalization. 00:15:20.340 --> 00:15:22.770 align:middle line:84% Yawelmani is trying to avoid three consonants in a row. 00:15:22.770 --> 00:15:27.360 align:middle line:84% Japanese is trying to avoid two accents in a single word. 00:15:27.360 --> 00:15:28.230 align:middle line:90% Maybe it's obvious. 00:15:28.230 --> 00:15:30.520 align:middle line:84% We don't want to miss generalizations like that. 00:15:30.520 --> 00:15:34.860 align:middle line:84% We want to find them and understand them. 00:15:34.860 --> 00:15:38.880 align:middle line:84% So similar kind of thing that we could think about here-- 00:15:38.880 --> 00:15:41.610 align:middle line:84% we've talked now a couple of times about Polish. 00:15:41.610 --> 00:15:43.530 align:middle line:84% Polish is one of the number of languages 00:15:43.530 --> 00:15:47.040 align:middle line:84% in which if you have certain kinds of consonants 00:15:47.040 --> 00:15:49.350 align:middle line:84% at the ends of words, consonants like "b," 00:15:49.350 --> 00:15:51.750 align:middle line:84% and "d," and "g," and "z," we convince ourselves 00:15:51.750 --> 00:15:52.560 align:middle line:90% that they devoice. 00:15:52.560 --> 00:15:55.170 align:middle line:84% They become voiceless at the ends of words. 00:15:55.170 --> 00:15:57.780 align:middle line:84% That allows us to capture various kinds of facts 00:15:57.780 --> 00:16:00.720 align:middle line:84% about what happens to Polish words when they become plural-- 00:16:00.720 --> 00:16:06.000 align:middle line:84% namely, that consonants that sound voiceless when they're 00:16:06.000 --> 00:16:09.240 align:middle line:84% at the ends of words become voiced sometimes 00:16:09.240 --> 00:16:11.640 align:middle line:84% when they're no longer at the end of words, we decided. 00:16:11.640 --> 00:16:14.280 align:middle line:84% That's because they're underlyingly voiced. 00:16:14.280 --> 00:16:18.730 align:middle line:84% And they devoice when they're at the ends of words. 00:16:18.730 --> 00:16:20.280 align:middle line:90% So this is all familiar. 00:16:20.280 --> 00:16:21.745 align:middle line:90% You all remember Polish? 00:16:21.745 --> 00:16:22.620 align:middle line:90% It's not just Polish. 00:16:22.620 --> 00:16:26.350 align:middle line:84% There are lots of languages that do this. 00:16:26.350 --> 00:16:30.030 align:middle line:84% And so what I'm inviting you to do at the end of this slide 00:16:30.030 --> 00:16:33.180 align:middle line:84% is to imagine two kinds of stories about Polish-- one 00:16:33.180 --> 00:16:36.360 align:middle line:84% that would say, yeah, final consonants devoice in Polish, 00:16:36.360 --> 00:16:38.370 align:middle line:84% and another that would say, well, there 00:16:38.370 --> 00:16:39.480 align:middle line:90% are four rules, actually. 00:16:39.480 --> 00:16:42.210 align:middle line:84% There's a rule that says "b" voices when it's final, 00:16:42.210 --> 00:16:44.520 align:middle line:84% and a rule that says "d" devoices when it's final, 00:16:44.520 --> 00:16:47.790 align:middle line:84% and a rule that says "g" devoices when it's-- yeah? 00:16:47.790 --> 00:16:50.640 align:middle line:84% Both of those would capture all of the Polish facts. 00:16:50.640 --> 00:16:53.230 align:middle line:90% One of them is clearly better. 00:16:53.230 --> 00:16:57.660 align:middle line:84% So we want to understand why we get the particular phenomena 00:16:57.660 --> 00:16:58.890 align:middle line:90% that we do in Polish. 00:16:58.890 --> 00:17:01.470 align:middle line:84% Is it clear to everybody that it would be better 00:17:01.470 --> 00:17:03.310 align:middle line:90% to do things that way? 00:17:03.310 --> 00:17:07.630 align:middle line:84% And what we're doing here is similar. 00:17:07.630 --> 00:17:10.920 align:middle line:90% Why would it be better? 00:17:10.920 --> 00:17:12.230 align:middle line:90% What would be wrong-- 00:17:12.230 --> 00:17:15.599 align:middle line:84% why is it better to be able to say final consonants devoice 00:17:15.599 --> 00:17:19.900 align:middle line:84% than to say, well, we have these four rules. 00:17:19.900 --> 00:17:20.490 align:middle line:90% Yeah, Faith? 00:17:20.490 --> 00:17:21.660 align:middle line:84% STUDENT: I think you're addressing 00:17:21.660 --> 00:17:22.948 align:middle line:90% the more general phenomenon. 00:17:22.948 --> 00:17:25.852 align:middle line:84% It's very difficult to catalog every single letter 00:17:25.852 --> 00:17:29.724 align:middle line:84% and to have vowel sound as something like that 00:17:29.724 --> 00:17:32.053 align:middle line:84% because you eventually have a list that's very long. 00:17:32.053 --> 00:17:32.970 align:middle line:90% NORVIN RICHARDS: Yeah. 00:17:32.970 --> 00:17:34.410 align:middle line:84% So it might make it easier to make 00:17:34.410 --> 00:17:35.970 align:middle line:90% cross-linguistic comparisons. 00:17:35.970 --> 00:17:37.350 align:middle line:90% Yeah, that's a very good point. 00:17:37.350 --> 00:17:39.930 align:middle line:90% Yeah, yeah. 00:17:39.930 --> 00:17:40.430 align:middle line:90% Yeah? 00:17:40.430 --> 00:17:43.260 align:middle line:90% 00:17:43.260 --> 00:17:45.285 align:middle line:90% We also maybe-- 00:17:45.285 --> 00:17:47.160 align:middle line:84% I don't know if this is a good way to say it. 00:17:47.160 --> 00:17:49.030 align:middle line:90% Maybe it is. 00:17:49.030 --> 00:17:53.650 align:middle line:84% We want to understand why the sounds that become 00:17:53.650 --> 00:17:56.845 align:middle line:84% voiceless at the ends of words in Polish are "b," 00:17:56.845 --> 00:18:04.690 align:middle line:84% "d," "g," and "z," and not "b," "d," "a," and "o," right? 00:18:04.690 --> 00:18:08.470 align:middle line:84% If we're just making four rules for those four letters, 00:18:08.470 --> 00:18:11.590 align:middle line:90% then it's not clear. 00:18:11.590 --> 00:18:14.320 align:middle line:84% There's no reason why one of those rules 00:18:14.320 --> 00:18:15.740 align:middle line:90% is more natural than the other. 00:18:15.740 --> 00:18:17.620 align:middle line:90% But as I keep saying-- 00:18:17.620 --> 00:18:20.270 align:middle line:84% what's happening here in Polish is extremely common. 00:18:20.270 --> 00:18:22.283 align:middle line:90% So you get it in Turkish. 00:18:22.283 --> 00:18:23.200 align:middle line:90% You get it in Russian. 00:18:23.200 --> 00:18:25.810 align:middle line:84% You get it in a bunch of different languages-- 00:18:25.810 --> 00:18:27.700 align:middle line:90% German. 00:18:27.700 --> 00:18:31.512 align:middle line:84% And so we want to understand why it's common. 00:18:31.512 --> 00:18:33.220 align:middle line:84% And part of understanding why it's common 00:18:33.220 --> 00:18:36.880 align:middle line:84% is being able to say that that's a common rule, a standard rule 00:18:36.880 --> 00:18:38.540 align:middle line:90% that we see all over the place. 00:18:38.540 --> 00:18:41.080 align:middle line:84% And then we can get to work on understanding it. 00:18:41.080 --> 00:18:42.520 align:middle line:84% If we just have these four rules, 00:18:42.520 --> 00:18:44.860 align:middle line:84% then it's unclear why those four rules are so common, 00:18:44.860 --> 00:18:49.720 align:middle line:84% whereas "b," "d," "a," and "o" devoice at the ends of words 00:18:49.720 --> 00:18:50.837 align:middle line:90% is not at all common. 00:18:50.837 --> 00:18:52.420 align:middle line:84% I don't know of any languages that do. 00:18:52.420 --> 00:18:54.130 align:middle line:90% Does that make sense? 00:18:54.130 --> 00:18:57.730 align:middle line:84% That's another sense in which making generalizations is good. 00:18:57.730 --> 00:19:00.430 align:middle line:84% This is becoming a philosophy of science question, yeah. 00:19:00.430 --> 00:19:02.560 align:middle line:90% It's like, why is science good? 00:19:02.560 --> 00:19:04.785 align:middle line:84% Why is it good to make generalizations? 00:19:04.785 --> 00:19:06.160 align:middle line:84% Because it helps us to understand 00:19:06.160 --> 00:19:09.075 align:middle line:84% the world better, for Faith's reason, as well as the reason 00:19:09.075 --> 00:19:09.700 align:middle line:90% I just offered. 00:19:09.700 --> 00:19:10.200 align:middle line:90% Yeah? 00:19:10.200 --> 00:19:14.368 align:middle line:84% STUDENT: Is the vowel still a vowel when you devoice it? 00:19:14.368 --> 00:19:15.910 align:middle line:84% NORVIN RICHARDS: Yeah, yes, arguably. 00:19:15.910 --> 00:19:17.830 align:middle line:84% So there are languages that devoice vowels. 00:19:17.830 --> 00:19:20.800 align:middle line:90% 00:19:20.800 --> 00:19:22.660 align:middle line:90% Mokilese is one. 00:19:22.660 --> 00:19:24.520 align:middle line:90% Oh, Japanese is another. 00:19:24.520 --> 00:19:31.690 align:middle line:84% So Japanese has a general rule devoicing its high vowels 00:19:31.690 --> 00:19:36.550 align:middle line:84% when they are between voiceless consonants, I guess. 00:19:36.550 --> 00:19:41.260 align:middle line:84% So here's the Japanese word for "like." 00:19:41.260 --> 00:19:43.990 align:middle line:84% And it's pronounced "suki," "suki." 00:19:43.990 --> 00:19:45.640 align:middle line:90% So the oo isn't pronounced. 00:19:45.640 --> 00:19:48.820 align:middle line:84% You don't say "soo-kee" unless you're being very slow 00:19:48.820 --> 00:19:50.560 align:middle line:90% and careful. 00:19:50.560 --> 00:19:53.738 align:middle line:84% And it's still a vowel in the sense 00:19:53.738 --> 00:19:55.780 align:middle line:84% that your tongue and the rest of your vocal tract 00:19:55.780 --> 00:19:57.010 align:middle line:84% is still in the configuration that it 00:19:57.010 --> 00:19:58.930 align:middle line:84% would be if your vocal chords were vibrating. 00:19:58.930 --> 00:20:01.180 align:middle line:90% But they aren't. 00:20:01.180 --> 00:20:04.060 align:middle line:84% So yeah, there arguably are voiceless vowels. 00:20:04.060 --> 00:20:04.660 align:middle line:90% Yes, Joseph? 00:20:04.660 --> 00:20:06.910 align:middle line:84% STUDENT: What if you were putting an affix after that, 00:20:06.910 --> 00:20:08.380 align:middle line:90% which is "i" is high, too? 00:20:08.380 --> 00:20:09.940 align:middle line:90% So what would happen? 00:20:09.940 --> 00:20:11.080 align:middle line:90% NORVIN RICHARDS: Oh. 00:20:11.080 --> 00:20:12.700 align:middle line:84% So in this particular example, there 00:20:12.700 --> 00:20:17.800 align:middle line:84% aren't any affixes you can add to "suki" to devoice the vowel. 00:20:17.800 --> 00:20:23.620 align:middle line:90% But let me think of an example. 00:20:23.620 --> 00:20:26.620 align:middle line:90% 00:20:26.620 --> 00:20:27.120 align:middle line:90% Yeah. 00:20:27.120 --> 00:20:35.133 align:middle line:84% So "tatsu"-- so this is not high. 00:20:35.133 --> 00:20:36.050 align:middle line:90% So it doesn't devoice. 00:20:36.050 --> 00:20:39.140 align:middle line:90% This is the verb "to stand." 00:20:39.140 --> 00:20:43.500 align:middle line:84% The more polite version of this-- 00:20:43.500 --> 00:20:44.200 align:middle line:90% no, no, no, no. 00:20:44.200 --> 00:20:47.033 align:middle line:90% That's not what I want to do. 00:20:47.033 --> 00:20:48.450 align:middle line:84% Let me try to think of an example. 00:20:48.450 --> 00:20:50.033 align:middle line:84% Let me try to think of an example that 00:20:50.033 --> 00:20:51.180 align:middle line:90% does what you want. 00:20:51.180 --> 00:20:54.510 align:middle line:84% But there aren't any examples that will work with "suki." 00:20:54.510 --> 00:20:57.390 align:middle line:90% Yeah. 00:20:57.390 --> 00:20:59.505 align:middle line:84% OK, so when I think of an example, 00:20:59.505 --> 00:21:01.380 align:middle line:84% at 3:00 in the morning, I'll give you a call. 00:21:01.380 --> 00:21:05.430 align:middle line:84% Leave your phone number and think of that. 00:21:05.430 --> 00:21:07.570 align:middle line:84% There's got to be a straightforward way to do this. 00:21:07.570 --> 00:21:10.080 align:middle line:84% Hmm, this is going to bother me now. 00:21:10.080 --> 00:21:10.875 align:middle line:90% Yeah. 00:21:10.875 --> 00:21:12.000 align:middle line:90% Other questions about this? 00:21:12.000 --> 00:21:16.175 align:middle line:90% 00:21:16.175 --> 00:21:17.550 align:middle line:84% Try to think of questions while I 00:21:17.550 --> 00:21:20.460 align:middle line:84% try to think of examples that do what Joseph wants. 00:21:20.460 --> 00:21:24.030 align:middle line:84% Because the past tense suffix in Japanese-- oh, oh, oh. 00:21:24.030 --> 00:21:25.260 align:middle line:90% [JAPANESE] oh, yeah, OK. 00:21:25.260 --> 00:21:25.770 align:middle line:90% Here's one. 00:21:25.770 --> 00:21:28.650 align:middle line:90% 00:21:28.650 --> 00:21:31.980 align:middle line:90% Adjectives in Japanese. 00:21:31.980 --> 00:21:37.440 align:middle line:84% So this is the word for new, "atarashii." 00:21:37.440 --> 00:21:45.580 align:middle line:84% And the negative of it, "atarashiku nai" ends-- 00:21:45.580 --> 00:21:49.230 align:middle line:84% so there's a "ku" that you add to make the negative form. 00:21:49.230 --> 00:21:53.640 align:middle line:84% And it ends up putting this vowel between voiceless sounds. 00:21:53.640 --> 00:21:55.080 align:middle line:90% And then yes, it devoices. 00:21:55.080 --> 00:21:59.070 align:middle line:84% So it goes from "atarashii" to "atarashiku nai," 00:21:59.070 --> 00:22:03.030 align:middle line:84% "atarashiku nai," or the past tense, actually. 00:22:03.030 --> 00:22:04.710 align:middle line:84% If you want to say something was new, 00:22:04.710 --> 00:22:06.660 align:middle line:90% it's "atarashii" [JAPANESE]. 00:22:06.660 --> 00:22:09.730 align:middle line:84% So again, the "i" that's after the "sh" 00:22:09.730 --> 00:22:12.390 align:middle line:84% devoices because there's a voiceless sound after. 00:22:12.390 --> 00:22:13.300 align:middle line:90% Good. 00:22:13.300 --> 00:22:13.980 align:middle line:90% Phew! 00:22:13.980 --> 00:22:16.550 align:middle line:84% Now I won't call you at 2:00 in the morning. 00:22:16.550 --> 00:22:18.959 align:middle line:84% You have any questions about this? 00:22:18.959 --> 00:22:20.380 align:middle line:90% OK. 00:22:20.380 --> 00:22:22.730 align:middle line:90% OK. 00:22:22.730 --> 00:22:29.030 align:middle line:84% So let's go back now to the English plural suffix 00:22:29.030 --> 00:22:33.380 align:middle line:84% and review something that we have talked about. 00:22:33.380 --> 00:22:35.540 align:middle line:84% But in the light of all of the things that we've 00:22:35.540 --> 00:22:39.300 align:middle line:84% talked about since then, I hope it'll become a little clearer. 00:22:39.300 --> 00:22:42.650 align:middle line:84% So again, we're talking about the regular English plural 00:22:42.650 --> 00:22:43.310 align:middle line:90% suffix. 00:22:43.310 --> 00:22:46.670 align:middle line:84% Stop thinking about "oxen" and "fish" and "sheep." 00:22:46.670 --> 00:22:48.980 align:middle line:90% "Cats" and "dogs" and "bushes"-- 00:22:48.980 --> 00:22:51.230 align:middle line:84% we decided the underlying form of the plural suffix 00:22:51.230 --> 00:22:52.100 align:middle line:90% is a "z," right? 00:22:52.100 --> 00:22:56.060 align:middle line:84% And it undergoes various kinds of sound changes 00:22:56.060 --> 00:22:58.220 align:middle line:90% depending on what's before it. 00:22:58.220 --> 00:23:02.510 align:middle line:84% So you add "z" to all three of these underlyingly. 00:23:02.510 --> 00:23:05.240 align:middle line:84% And then there are generalizations like English 00:23:05.240 --> 00:23:07.520 align:middle line:84% doesn't like words that end in a "z" that's 00:23:07.520 --> 00:23:09.950 align:middle line:90% preceded by a voiceless sound. 00:23:09.950 --> 00:23:15.290 align:middle line:84% And words don't get to end in two strident consonants, where 00:23:15.290 --> 00:23:18.260 align:middle line:84% stridents are things like the "sh" and the "z" 00:23:18.260 --> 00:23:22.340 align:middle line:84% that's at the end of "bushes," yeah? 00:23:22.340 --> 00:23:23.990 align:middle line:90% So those are generalizations. 00:23:23.990 --> 00:23:26.065 align:middle line:84% But we said one of the attractive things 00:23:26.065 --> 00:23:28.190 align:middle line:84% about this way of thinking about the English plural 00:23:28.190 --> 00:23:29.565 align:middle line:84% is that these are generalizations 00:23:29.565 --> 00:23:31.760 align:middle line:84% that just seem to hold about English generally. 00:23:31.760 --> 00:23:33.290 align:middle line:84% English doesn't have any words that 00:23:33.290 --> 00:23:35.750 align:middle line:84% end in a "z" preceded by a voiceless sound. 00:23:35.750 --> 00:23:38.120 align:middle line:84% And we can rely on English phonology 00:23:38.120 --> 00:23:41.535 align:middle line:84% to demand that you do something if you add "z" to cat. 00:23:41.535 --> 00:23:44.480 align:middle line:90% 00:23:44.480 --> 00:23:48.020 align:middle line:84% And again, we've said this before. 00:23:48.020 --> 00:23:50.540 align:middle line:84% I just want to say it again a little more clearly. 00:23:50.540 --> 00:23:52.220 align:middle line:90% There are these conditions. 00:23:52.220 --> 00:23:54.560 align:middle line:84% And then there are also these procedures 00:23:54.560 --> 00:23:56.630 align:middle line:90% for repairing these violations. 00:23:56.630 --> 00:24:00.920 align:middle line:84% So what do you do about the fact that you add a "z" to cat? 00:24:00.920 --> 00:24:03.920 align:middle line:84% Well, you devoice the "z," right? 00:24:03.920 --> 00:24:06.470 align:middle line:84% If the condition just says you can't have a "z" preceded 00:24:06.470 --> 00:24:08.810 align:middle line:84% by something voiceless, there are various things 00:24:08.810 --> 00:24:10.790 align:middle line:84% you could have done to fix that problem. 00:24:10.790 --> 00:24:13.250 align:middle line:84% You could have voiced the "d," right? 00:24:13.250 --> 00:24:15.500 align:middle line:84% So the plural of "cat" could be "cads," right? 00:24:15.500 --> 00:24:18.080 align:middle line:84% That would be a good English word. 00:24:18.080 --> 00:24:21.470 align:middle line:84% Or you could add a vowel like you do in "bushes," right? 00:24:21.470 --> 00:24:23.420 align:middle line:84% The plural of "cat" could be "cat-es." 00:24:23.420 --> 00:24:27.140 align:middle line:84% There are various things that would fix that problem 00:24:27.140 --> 00:24:28.130 align:middle line:90% if that's the problem. 00:24:28.130 --> 00:24:29.960 align:middle line:84% But this is the thing that you do. 00:24:29.960 --> 00:24:32.360 align:middle line:84% Yeah, it's like when we were talking about Yawelmani 00:24:32.360 --> 00:24:34.470 align:middle line:84% We said you can't have three consonants in a row. 00:24:34.470 --> 00:24:35.550 align:middle line:90% What do you do about it? 00:24:35.550 --> 00:24:37.423 align:middle line:84% Well, it depends on the situation, right? 00:24:37.423 --> 00:24:39.590 align:middle line:84% So if one of them is an "h," you get rid of the "h." 00:24:39.590 --> 00:24:42.380 align:middle line:84% If not, then you insert a vowel in this position 00:24:42.380 --> 00:24:43.720 align:middle line:90% and not that position. 00:24:43.720 --> 00:24:47.700 align:middle line:90% It's similar thing. 00:24:47.700 --> 00:24:54.170 align:middle line:84% So this is all the way of saying it's worthwhile, maybe, 00:24:54.170 --> 00:24:58.670 align:middle line:84% to distinguish between the phonotactic considerations 00:24:58.670 --> 00:25:01.220 align:middle line:84% or the well-formedness considerations that tell you 00:25:01.220 --> 00:25:04.340 align:middle line:84% this language doesn't like three consonants in a row, 00:25:04.340 --> 00:25:06.320 align:middle line:84% or two accented syllables in a word, 00:25:06.320 --> 00:25:10.610 align:middle line:84% or words ending in a "z" preceded by a voiceless sound. 00:25:10.610 --> 00:25:12.200 align:middle line:90% There's that condition. 00:25:12.200 --> 00:25:14.900 align:middle line:84% And then there's how do you repair it? 00:25:14.900 --> 00:25:17.270 align:middle line:84% What kind of a repair do you use? 00:25:17.270 --> 00:25:19.460 align:middle line:84% So there's the well-formedness condition. 00:25:19.460 --> 00:25:20.820 align:middle line:90% And then there's the repair. 00:25:20.820 --> 00:25:22.820 align:middle line:84% So in this particular case, the procedures 00:25:22.820 --> 00:25:26.990 align:middle line:90% are devoice and insert schwa. 00:25:26.990 --> 00:25:32.750 align:middle line:84% There's a particularly striking bit of evidence for this way 00:25:32.750 --> 00:25:36.080 align:middle line:84% of dividing things up, this distinction 00:25:36.080 --> 00:25:38.870 align:middle line:84% that I'm drawing between well-formedness conditions 00:25:38.870 --> 00:25:43.400 align:middle line:84% on words, on the one hand, and types of possible repair 00:25:43.400 --> 00:25:44.660 align:middle line:90% on the other. 00:25:44.660 --> 00:25:46.550 align:middle line:84% We need to think about both of those things 00:25:46.550 --> 00:25:50.120 align:middle line:84% if we're going to describe what the language is doing. 00:25:50.120 --> 00:25:53.180 align:middle line:84% This is a classic test from 1957. 00:25:53.180 --> 00:25:54.733 align:middle line:84% I mentioned it in class before, but I 00:25:54.733 --> 00:25:56.900 align:middle line:84% wanted to talk about it a little more now that we've 00:25:56.900 --> 00:25:58.790 align:middle line:90% drawn this distinction. 00:25:58.790 --> 00:26:04.070 align:middle line:84% So Jean Berko, later Jean Berko Gleason, 00:26:04.070 --> 00:26:07.700 align:middle line:84% did this work in 1957 with children. 00:26:07.700 --> 00:26:11.360 align:middle line:84% She was demonstrating that they know these generalizations 00:26:11.360 --> 00:26:15.420 align:middle line:84% about plurals even with words they've never heard before. 00:26:15.420 --> 00:26:17.453 align:middle line:84% So she would show them pictures like this one. 00:26:17.453 --> 00:26:18.620 align:middle line:90% This is one of her pictures. 00:26:18.620 --> 00:26:20.900 align:middle line:90% She said "This is a wug." 00:26:20.900 --> 00:26:25.430 align:middle line:84% And then she would show them two pictures of the same thing. 00:26:25.430 --> 00:26:28.040 align:middle line:84% And she would say, "Now there are two of them. 00:26:28.040 --> 00:26:35.840 align:middle line:84% There are two"-- and children were supposed to say "wugs" 00:26:35.840 --> 00:26:38.410 align:middle line:84% even though they'd never heard anybody say "wug" before, 00:26:38.410 --> 00:26:38.910 align:middle line:90% right? 00:26:38.910 --> 00:26:42.680 align:middle line:84% So Jean Berko Gleason, Jean Berko made up a bunch of words. 00:26:42.680 --> 00:26:45.110 align:middle line:84% She worked hard to make sure they were not real words. 00:26:45.110 --> 00:26:47.660 align:middle line:84% And then she drew many cute pictures 00:26:47.660 --> 00:26:50.650 align:middle line:84% and showed children singulars and plurals and said, 00:26:50.650 --> 00:26:51.650 align:middle line:90% OK, here's the singular. 00:26:51.650 --> 00:26:52.790 align:middle line:90% What's the plural? 00:26:52.790 --> 00:26:55.352 align:middle line:90% And the kids responded. 00:26:55.352 --> 00:26:56.810 align:middle line:84% And she wrote down their responses. 00:26:56.810 --> 00:26:58.790 align:middle line:84% And their responses, interestingly, 00:26:58.790 --> 00:27:02.930 align:middle line:84% were accurate in one way, and sometimes inaccurate 00:27:02.930 --> 00:27:05.187 align:middle line:84% in another, which is interesting to think about. 00:27:05.187 --> 00:27:07.520 align:middle line:84% So she was working with children who were five years old 00:27:07.520 --> 00:27:09.483 align:middle line:84% and children who were seven years old. 00:27:09.483 --> 00:27:11.150 align:middle line:84% What she discovered was that by the time 00:27:11.150 --> 00:27:13.108 align:middle line:84% they were five years old, they are, in a sense, 00:27:13.108 --> 00:27:17.660 align:middle line:84% perfect, in the sense that all of the words that they came up 00:27:17.660 --> 00:27:21.600 align:middle line:84% with were phonotactically well formed English words. 00:27:21.600 --> 00:27:24.860 align:middle line:84% So here are three of Berko's words-- "hif," 00:27:24.860 --> 00:27:27.370 align:middle line:90% "fas," and "muz," yeah? 00:27:27.370 --> 00:27:31.100 align:middle line:84% And I think everybody did "wug" perfectly. 00:27:31.100 --> 00:27:33.830 align:middle line:84% Everybody said the plural of "wug" is "wugs." 00:27:33.830 --> 00:27:37.310 align:middle line:84% But there were other cases where children would say, 00:27:37.310 --> 00:27:40.820 align:middle line:84% sometimes, things like the plural of "hif" is "hifes," 00:27:40.820 --> 00:27:44.540 align:middle line:84% or the plural of "fas" is "fas," or the plural of "muz" 00:27:44.540 --> 00:27:47.740 align:middle line:90% is "muz." 00:27:47.740 --> 00:27:50.178 align:middle line:90% These are all words where-- 00:27:50.178 --> 00:27:51.970 align:middle line:84% and then there were other kinds of mistakes 00:27:51.970 --> 00:27:53.110 align:middle line:90% that children didn't make. 00:27:53.110 --> 00:27:58.000 align:middle line:84% So nobody said that the plural "wug" was "wuks," 00:27:58.000 --> 00:28:02.830 align:middle line:84% or that the plural of "wook" is "wookz," right? 00:28:02.830 --> 00:28:07.630 align:middle line:84% Or that the plural of fas is "fas-z," yeah? 00:28:07.630 --> 00:28:08.830 align:middle line:90% What were children doing? 00:28:08.830 --> 00:28:11.170 align:middle line:84% Well, children know if you show them 00:28:11.170 --> 00:28:15.010 align:middle line:84% something that's underlyingly-- you give them "hif," which 00:28:15.010 --> 00:28:17.595 align:middle line:84% for some reason I'm spelling backwards-- you give them "hif" 00:28:17.595 --> 00:28:18.970 align:middle line:84% and you ask them the plural, they 00:28:18.970 --> 00:28:21.610 align:middle line:84% know that you're supposed to add a "z." 00:28:21.610 --> 00:28:24.220 align:middle line:84% And they know that English doesn't allow words 00:28:24.220 --> 00:28:28.000 align:middle line:84% that end in a "z" preceded by a voiceless sound. 00:28:28.000 --> 00:28:29.920 align:middle line:84% But some of these five and seven-year-olds 00:28:29.920 --> 00:28:32.290 align:middle line:84% were a little fuzzy on the possible repairs. 00:28:32.290 --> 00:28:35.290 align:middle line:90% 00:28:35.290 --> 00:28:38.740 align:middle line:84% None of them said, "Oh, yeah, It's hifz," right? 00:28:38.740 --> 00:28:40.930 align:middle line:90% But I just said this. 00:28:40.930 --> 00:28:43.420 align:middle line:84% If the only problem is, well, this word 00:28:43.420 --> 00:28:44.920 align:middle line:84% ends in a voiceless sound followed 00:28:44.920 --> 00:28:47.590 align:middle line:84% by a "z," in fact, what you're supposed 00:28:47.590 --> 00:28:49.360 align:middle line:90% to do is devoice the "z." 00:28:49.360 --> 00:28:50.890 align:middle line:90% That's what an adult would do. 00:28:50.890 --> 00:28:53.290 align:middle line:84% So the plural of "hif" is "hifs," yeah? 00:28:53.290 --> 00:28:57.730 align:middle line:84% But if you were instead to insert a schwa, 00:28:57.730 --> 00:28:59.860 align:middle line:84% well, that would fix it, too, right? 00:28:59.860 --> 00:29:03.670 align:middle line:84% So now the "z" is preceded by a voiced sound. 00:29:03.670 --> 00:29:05.322 align:middle line:84% The plural of "hif" is "hif-ez," yeah? 00:29:05.322 --> 00:29:07.780 align:middle line:84% And there were some kids who were going through a stage who 00:29:07.780 --> 00:29:10.180 align:middle line:84% were entertaining that possibility, that that was 00:29:10.180 --> 00:29:12.160 align:middle line:90% the correct repair for that. 00:29:12.160 --> 00:29:18.160 align:middle line:84% So what Berko found was that children universally 00:29:18.160 --> 00:29:20.200 align:middle line:90% knew the phonotactics, right? 00:29:20.200 --> 00:29:23.500 align:middle line:84% They knew that English doesn't allow words 00:29:23.500 --> 00:29:25.930 align:middle line:84% that end in a "z" preceded by a voiceless sound. 00:29:25.930 --> 00:29:31.150 align:middle line:84% But they made intriguing mistakes of various kinds-- 00:29:31.150 --> 00:29:32.192 align:middle line:90% not very often. 00:29:32.192 --> 00:29:34.150 align:middle line:84% Most of the kids just did everything perfectly. 00:29:34.150 --> 00:29:35.140 align:middle line:84% But every so often, she would get 00:29:35.140 --> 00:29:36.340 align:middle line:90% kids who made some mistakes. 00:29:36.340 --> 00:29:38.560 align:middle line:84% And the mistakes were of this class. 00:29:38.560 --> 00:29:41.590 align:middle line:84% That is, mistakes where they knew 00:29:41.590 --> 00:29:45.110 align:middle line:84% that there needed to be a repair but got the repair wrong. 00:29:45.110 --> 00:29:48.970 align:middle line:84% Similarly for "muz," this was a kid who-- 00:29:48.970 --> 00:29:52.570 align:middle line:84% the singular is "muz," and the kid knows 00:29:52.570 --> 00:29:53.830 align:middle line:90% that you need to add a "z." 00:29:53.830 --> 00:29:55.955 align:middle line:84% And the kid knows that you can't have two stridents 00:29:55.955 --> 00:29:56.980 align:middle line:90% at the end of a word. 00:29:56.980 --> 00:29:59.140 align:middle line:84% The correct answer, the adult's answer, 00:29:59.140 --> 00:30:01.810 align:middle line:84% is to insert a schwa so the plural of "muz" 00:30:01.810 --> 00:30:02.770 align:middle line:90% will be "muzes." 00:30:02.770 --> 00:30:08.050 align:middle line:84% But deleting the "z," one of the "z"s-- hard to know which one-- 00:30:08.050 --> 00:30:09.590 align:middle line:90% also fixes that problem. 00:30:09.590 --> 00:30:11.800 align:middle line:84% And so that was a kid who thought, 00:30:11.800 --> 00:30:13.720 align:middle line:90% yeah, I'll delete a "z," yeah? 00:30:13.720 --> 00:30:14.938 align:middle line:90% Yeah, Joseph? 00:30:14.938 --> 00:30:17.900 align:middle line:84% STUDENT: Is the deletion of a consonant viewed 00:30:17.900 --> 00:30:19.260 align:middle line:90% as [INAUDIBLE]? 00:30:19.260 --> 00:30:21.260 align:middle line:84% NORVIN RICHARDS: As something that English does? 00:30:21.260 --> 00:30:22.010 align:middle line:90% Yeah, no. 00:30:22.010 --> 00:30:24.020 align:middle line:84% So English-- there aren't, I don't think, 00:30:24.020 --> 00:30:30.050 align:middle line:84% any English nouns that form their plural by adding a "z" 00:30:30.050 --> 00:30:31.795 align:middle line:84% and deleting something before the "z." 00:30:31.795 --> 00:30:34.880 align:middle line:84% I don't believe that ever happens. 00:30:34.880 --> 00:30:36.860 align:middle line:84% This is me slowing down while I try 00:30:36.860 --> 00:30:39.410 align:middle line:84% to think of English nouns, of which I know several, 00:30:39.410 --> 00:30:42.710 align:middle line:84% but I can't think of any that do that. 00:30:42.710 --> 00:30:43.730 align:middle line:90% Yeah. 00:30:43.730 --> 00:30:44.960 align:middle line:90% Yeah. 00:30:44.960 --> 00:30:46.490 align:middle line:90% OK. 00:30:46.490 --> 00:30:48.230 align:middle line:90% Yeah, so is this clear? 00:30:48.230 --> 00:30:49.970 align:middle line:84% Another reason to draw this distinction 00:30:49.970 --> 00:30:52.970 align:middle line:84% between, on the one hand, what I was calling phonotactics, 00:30:52.970 --> 00:30:56.010 align:middle line:84% these rules for which kinds of sounds can be where, 00:30:56.010 --> 00:30:57.900 align:middle line:84% and on the other hand, the repairs-- 00:30:57.900 --> 00:31:01.130 align:middle line:84% so the kids that Berko was working with 00:31:01.130 --> 00:31:04.100 align:middle line:84% knew the phonotactics, but were sometimes fuzzy on what 00:31:04.100 --> 00:31:06.683 align:middle line:90% the correct repairs were. 00:31:06.683 --> 00:31:08.570 align:middle line:90% Yeah. 00:31:08.570 --> 00:31:10.220 align:middle line:90% OK. 00:31:10.220 --> 00:31:12.853 align:middle line:84% So, moral-- yet again, need to distinguish between, 00:31:12.853 --> 00:31:14.270 align:middle line:84% on the one hand, the phonotactics, 00:31:14.270 --> 00:31:16.670 align:middle line:84% and on the other hand, the particular sound changes, 00:31:16.670 --> 00:31:18.630 align:middle line:84% the repairs that enforce the phonotactics. 00:31:18.630 --> 00:31:24.160 align:middle line:84% It's apparently possible to know one of those but not the other. 00:31:24.160 --> 00:31:25.200 align:middle line:90% OK. 00:31:25.200 --> 00:31:28.370 align:middle line:90% That ends the chapter. 00:31:28.370 --> 00:31:31.337 align:middle line:84% Now we're going to talk about something else. 00:31:31.337 --> 00:31:32.920 align:middle line:84% So are there any questions about that? 00:31:32.920 --> 00:31:34.132 align:middle line:90% Is that clear? 00:31:34.132 --> 00:31:35.980 align:middle line:90% OK. 00:31:35.980 --> 00:31:38.980 align:middle line:84% We've already done some of this, but I want to do more. 00:31:38.980 --> 00:31:42.730 align:middle line:84% We've been talking about how there 00:31:42.730 --> 00:31:46.840 align:middle line:84% are some kinds of sound changes that are natural because they 00:31:46.840 --> 00:31:48.940 align:middle line:84% apply to natural classes of things 00:31:48.940 --> 00:31:53.500 align:middle line:90% that can be almost circular. 00:31:53.500 --> 00:31:57.910 align:middle line:84% So I quickly introduced you to the word stridents 00:31:57.910 --> 00:32:00.700 align:middle line:84% because the generalization about English plurals 00:32:00.700 --> 00:32:04.210 align:middle line:84% is that two stridents can't end a word in English. 00:32:04.210 --> 00:32:07.060 align:middle line:84% And that's why you introduce a schwa just in case 00:32:07.060 --> 00:32:08.830 align:middle line:84% you're in danger of having two stridents 00:32:08.830 --> 00:32:12.370 align:middle line:90% at the end of your word, yeah. 00:32:12.370 --> 00:32:15.850 align:middle line:84% The word strident has an acoustic definition. 00:32:15.850 --> 00:32:19.420 align:middle line:84% They are fricatives that have lots of high frequency noise. 00:32:19.420 --> 00:32:22.990 align:middle line:84% But it's of interest that there are phenomena 00:32:22.990 --> 00:32:25.480 align:middle line:84% like the allomorphy in English plurals 00:32:25.480 --> 00:32:28.030 align:middle line:84% that make reference to that acoustic fact. 00:32:28.030 --> 00:32:31.720 align:middle line:84% That's apparently something that a language can care about-- 00:32:31.720 --> 00:32:33.983 align:middle line:84% how much high frequency noise you have at once. 00:32:33.983 --> 00:32:36.400 align:middle line:84% It's also something that varies from language to language. 00:32:36.400 --> 00:32:38.830 align:middle line:84% There are languages that are perfectly happy to have 00:32:38.830 --> 00:32:40.690 align:middle line:90% long strings of stridents. 00:32:40.690 --> 00:32:42.940 align:middle line:90% English just isn't one of them. 00:32:42.940 --> 00:32:45.430 align:middle line:90% Yeah. 00:32:45.430 --> 00:32:49.050 align:middle line:84% So we've already done, as I said, some talking about this. 00:32:49.050 --> 00:32:51.160 align:middle line:84% So when I was first introducing you to consonants 00:32:51.160 --> 00:32:53.470 align:middle line:84% or introducing you to vowels, you'd 00:32:53.470 --> 00:32:55.570 align:middle line:84% already met consonants at some point in your life. 00:32:55.570 --> 00:32:59.170 align:middle line:84% But we were talking about them in a particular systematic way. 00:32:59.170 --> 00:33:04.060 align:middle line:84% We talked about various ways of classifying them, saying 00:33:04.060 --> 00:33:06.340 align:middle line:84% there are some consonants that are fricatives 00:33:06.340 --> 00:33:08.020 align:middle line:84% and some consonants that are stops, 00:33:08.020 --> 00:33:10.210 align:middle line:84% or saying there are some consonants that are voiced 00:33:10.210 --> 00:33:12.410 align:middle line:84% and some consonants that are voiceless. 00:33:12.410 --> 00:33:14.920 align:middle line:84% So what I want to do today is to introduce you 00:33:14.920 --> 00:33:16.930 align:middle line:84% to some other kinds of classification. 00:33:16.930 --> 00:33:18.663 align:middle line:84% And I'll, as much as possible, try 00:33:18.663 --> 00:33:20.080 align:middle line:84% to show you places where languages 00:33:20.080 --> 00:33:23.500 align:middle line:84% care about a particular class of sounds. 00:33:23.500 --> 00:33:27.340 align:middle line:84% Again, this is all in the service of making it easier 00:33:27.340 --> 00:33:30.190 align:middle line:84% for us to understand why certain kinds of sound changes 00:33:30.190 --> 00:33:35.320 align:middle line:90% are natural and others are not. 00:33:35.320 --> 00:33:35.820 align:middle line:90% OK. 00:33:35.820 --> 00:33:39.840 align:middle line:84% So here's one that we've already seen. 00:33:39.840 --> 00:33:41.175 align:middle line:90% There are sounds that are nasal. 00:33:41.175 --> 00:33:43.880 align:middle line:90% 00:33:43.880 --> 00:33:46.085 align:middle line:84% This is a contrast that I believe 00:33:46.085 --> 00:33:48.710 align:middle line:84% I have introduced you to before, but let me introduce you to it 00:33:48.710 --> 00:33:49.280 align:middle line:90% again. 00:33:49.280 --> 00:33:51.650 align:middle line:90% There's a distinction-- 00:33:51.650 --> 00:33:53.750 align:middle line:84% I think I brought this up last time, actually. 00:33:53.750 --> 00:33:56.780 align:middle line:84% There's a distinction, when we talk about consonants, 00:33:56.780 --> 00:33:59.660 align:middle line:84% between what are called sonorants 00:33:59.660 --> 00:34:01.650 align:middle line:90% and what are called obstruents. 00:34:01.650 --> 00:34:03.530 align:middle line:84% So these are opposites of each other. 00:34:03.530 --> 00:34:06.540 align:middle line:84% I hope I have the word "obstruent" on a later slide. 00:34:06.540 --> 00:34:10.580 align:middle line:84% But just in case I don't, there are these two kinds of sounds-- 00:34:10.580 --> 00:34:14.840 align:middle line:90% obstruents and sonorants. 00:34:14.840 --> 00:34:16.820 align:middle line:84% And the distinction between them has 00:34:16.820 --> 00:34:19.699 align:middle line:84% to do with whether you are building up 00:34:19.699 --> 00:34:22.110 align:middle line:90% pressure in the oral cavity. 00:34:22.110 --> 00:34:26.989 align:middle line:84% So the sonorants are the nasals and the liquids 00:34:26.989 --> 00:34:29.929 align:middle line:84% and the glides, the consonants in which there's 00:34:29.929 --> 00:34:33.139 align:middle line:84% no pressure building up in the oral cavity, 00:34:33.139 --> 00:34:35.150 align:middle line:84% whereas the obstruents-- those are 00:34:35.150 --> 00:34:38.690 align:middle line:84% the stops and the fricatives, the oral stops and fricatives. 00:34:38.690 --> 00:34:41.270 align:middle line:84% Those are the sounds in which there is pressure building up 00:34:41.270 --> 00:34:42.260 align:middle line:90% in the oral cavity. 00:34:42.260 --> 00:34:45.260 align:middle line:84% For a stop, it's because, well, you've stopped the flow of air 00:34:45.260 --> 00:34:46.159 align:middle line:90% at the mouth. 00:34:46.159 --> 00:34:47.719 align:middle line:90% And so pressure is building up. 00:34:47.719 --> 00:34:49.302 align:middle line:84% And for a fricative, well, you haven't 00:34:49.302 --> 00:34:51.469 align:middle line:84% stopped the flow of air, but you're constricting it. 00:34:51.469 --> 00:34:53.000 align:middle line:90% That's how you make a fricative. 00:34:53.000 --> 00:34:55.469 align:middle line:84% And so pressure is building up in the oral cavity. 00:34:55.469 --> 00:34:58.190 align:middle line:84% And when I say pressure is building up in the oral cavity, 00:34:58.190 --> 00:35:02.180 align:middle line:84% what I mean is that phoneticists have put instruments 00:35:02.180 --> 00:35:05.600 align:middle line:84% inside people's mouths to measure air pressure to find 00:35:05.600 --> 00:35:06.440 align:middle line:90% out what happens. 00:35:06.440 --> 00:35:09.730 align:middle line:84% And they have discovered that this is true, yeah. 00:35:09.730 --> 00:35:11.963 align:middle line:84% So I try not to lie to you in this class 00:35:11.963 --> 00:35:13.130 align:middle line:90% any more than I can help it. 00:35:13.130 --> 00:35:14.990 align:middle line:84% But in this particular case, what I'm saying 00:35:14.990 --> 00:35:15.890 align:middle line:90% is absolutely true. 00:35:15.890 --> 00:35:19.285 align:middle line:90% 00:35:19.285 --> 00:35:22.100 align:middle line:84% OK, so there are sonorants and there are obstruents. 00:35:22.100 --> 00:35:24.500 align:middle line:84% We have already talked about the classification of sounds 00:35:24.500 --> 00:35:25.910 align:middle line:84% into voiced and voiceless sounds. 00:35:25.910 --> 00:35:28.760 align:middle line:84% So the red sounds on this slide are voiced. 00:35:28.760 --> 00:35:30.620 align:middle line:84% And as you can see, in this cross-cut-- some 00:35:30.620 --> 00:35:31.828 align:middle line:90% of the other classifications. 00:35:31.828 --> 00:35:37.430 align:middle line:84% So you can be an obstruent, like a stop or a fricative, 00:35:37.430 --> 00:35:40.140 align:middle line:84% and you can be either voiced or voiceless. 00:35:40.140 --> 00:35:43.460 align:middle line:84% So English stops can be voiced or voiceless. 00:35:43.460 --> 00:35:46.220 align:middle line:84% English fricatives can be voiced or voiceless. 00:35:46.220 --> 00:35:48.110 align:middle line:84% English sonorants, interestingly, 00:35:48.110 --> 00:35:49.430 align:middle line:90% can only be voiced. 00:35:49.430 --> 00:35:51.410 align:middle line:90% They cannot be voiceless. 00:35:51.410 --> 00:35:54.830 align:middle line:84% We have sonorants like "l," like "l," 00:35:54.830 --> 00:35:56.865 align:middle line:84% but we don't have a voiceless version of "l." 00:35:56.865 --> 00:35:58.490 align:middle line:84% We talked about this when we were first 00:35:58.490 --> 00:36:00.253 align:middle line:90% going through sounds. 00:36:00.253 --> 00:36:01.670 align:middle line:84% There are languages out there that 00:36:01.670 --> 00:36:05.570 align:middle line:84% have a voiceless version of "l," that have a "lh" sound. 00:36:05.570 --> 00:36:07.320 align:middle line:90% Tibetan has one, for example. 00:36:07.320 --> 00:36:10.280 align:middle line:84% So if you go to Lhasa, the capital of Tibet, 00:36:10.280 --> 00:36:13.490 align:middle line:84% you must learn to make voiceless "l"s. 00:36:13.490 --> 00:36:16.760 align:middle line:90% Welsh has a voiceless "l.". 00:36:16.760 --> 00:36:18.560 align:middle line:90% They spell it with two "l"s. 00:36:18.560 --> 00:36:21.770 align:middle line:84% So if you look at Welsh words, you'll see pairs of "l"s. 00:36:21.770 --> 00:36:25.250 align:middle line:84% That's their way of spelling a voiceless "l." 00:36:25.250 --> 00:36:29.810 align:middle line:84% So there's a traditional Welsh name spelled like this 00:36:29.810 --> 00:36:32.660 align:middle line:84% and pronounced "Lhoyd," which has 00:36:32.660 --> 00:36:37.070 align:middle line:84% been borrowed into English twice as the name Lloyd and the name 00:36:37.070 --> 00:36:39.110 align:middle line:90% Floyd. 00:36:39.110 --> 00:36:40.637 align:middle line:84% These are both attempts to pronounce 00:36:40.637 --> 00:36:42.470 align:middle line:84% "Lhoyd," which we can't pronounce because we 00:36:42.470 --> 00:36:45.270 align:middle line:90% don't have that sound. 00:36:45.270 --> 00:36:45.770 align:middle line:90% Yeah. 00:36:45.770 --> 00:36:50.065 align:middle line:84% So English has voiced and voiceless obstruents, stops 00:36:50.065 --> 00:36:50.690 align:middle line:90% and fricatives. 00:36:50.690 --> 00:36:52.490 align:middle line:90% But it only has voice sonorants. 00:36:52.490 --> 00:36:55.160 align:middle line:84% That's a very common state of affairs. 00:36:55.160 --> 00:36:56.930 align:middle line:84% If you're going to have a voicing 00:36:56.930 --> 00:37:00.770 align:middle line:84% distinction in only one of these places, it's in the obstruents. 00:37:00.770 --> 00:37:02.660 align:middle line:84% Languages only having voiced sonorants-- 00:37:02.660 --> 00:37:06.860 align:middle line:84% that's very common, though not universal, as we've said. 00:37:06.860 --> 00:37:09.540 align:middle line:90% 00:37:09.540 --> 00:37:13.670 align:middle line:84% Quick interlude to use one of these features. 00:37:13.670 --> 00:37:18.510 align:middle line:84% Here are a bunch of verbs that you can make out of adjectives. 00:37:18.510 --> 00:37:21.410 align:middle line:84% So you can take an adjective like "black." 00:37:21.410 --> 00:37:23.867 align:middle line:84% And you can make a verb, "blacken," 00:37:23.867 --> 00:37:26.450 align:middle line:84% which can be either a transitive verb or in intransitive verb, 00:37:26.450 --> 00:37:26.950 align:middle line:90% I guess. 00:37:26.950 --> 00:37:33.650 align:middle line:84% So you can blacken the paper or blacken the fish, I guess. 00:37:33.650 --> 00:37:37.190 align:middle line:90% It means to make it black. 00:37:37.190 --> 00:37:43.190 align:middle line:84% Similarly, something can whiten or lessen or freshen or darken. 00:37:43.190 --> 00:37:45.230 align:middle line:84% You can offer to freshen a drink, say. 00:37:45.230 --> 00:37:48.020 align:middle line:84% It means to make it fresh, right? 00:37:48.020 --> 00:37:51.290 align:middle line:84% So there's this "-en" that you can add to adjectives to make 00:37:51.290 --> 00:37:51.920 align:middle line:90% verbs. 00:37:51.920 --> 00:37:54.870 align:middle line:90% Does that sound right? 00:37:54.870 --> 00:37:59.570 align:middle line:84% On the other hand, there are these other adjectives which 00:37:59.570 --> 00:38:01.250 align:middle line:90% can't have "-en," yeah? 00:38:01.250 --> 00:38:07.790 align:middle line:84% So the screen of a computer can darken, 00:38:07.790 --> 00:38:11.310 align:middle line:90% but it cannot "yellowen," yeah? 00:38:11.310 --> 00:38:14.390 align:middle line:90% A piece of paper that gets old-- 00:38:14.390 --> 00:38:16.160 align:middle line:84% what you say is that the yellows. 00:38:16.160 --> 00:38:17.540 align:middle line:90% It yellowed with age. 00:38:17.540 --> 00:38:20.000 align:middle line:84% You don't say that it "yellowened," yeah? 00:38:20.000 --> 00:38:22.670 align:middle line:84% Similarly, you can't "dimmen" or "grayen" 00:38:22.670 --> 00:38:24.593 align:middle line:90% or "clearen" or "brownen." 00:38:24.593 --> 00:38:26.510 align:middle line:84% All of those things-- if you want to make them 00:38:26.510 --> 00:38:27.860 align:middle line:90% into verbs, you just [SNAPS]. 00:38:27.860 --> 00:38:30.010 align:middle line:90% They're verbs, yeah? 00:38:30.010 --> 00:38:30.670 align:middle line:90% Yeah? 00:38:30.670 --> 00:38:32.770 align:middle line:90% Does that sound right? 00:38:32.770 --> 00:38:34.840 align:middle line:84% What's the difference between "black," 00:38:34.840 --> 00:38:38.980 align:middle line:84% "white," "less," "fresh," and "dark" on the one hand, 00:38:38.980 --> 00:38:43.090 align:middle line:84% and "yellow," "dim," "gray," "clear," 00:38:43.090 --> 00:38:46.843 align:middle line:90% and "brown" on the other? 00:38:46.843 --> 00:38:52.402 align:middle line:84% STUDENT: They end with voiceless stops in the first line? 00:38:52.402 --> 00:38:53.860 align:middle line:84% NORVIN RICHARDS: In the first line, 00:38:53.860 --> 00:38:57.490 align:middle line:84% "black," "white," and "dark" end with voiceless stops. 00:38:57.490 --> 00:39:02.190 align:middle line:84% And "less" and "fresh" end with voiceless fricatives, yeah? 00:39:02.190 --> 00:39:03.000 align:middle line:90% Yeah, yeah. 00:39:03.000 --> 00:39:06.340 align:middle line:90% Yeah, nice. 00:39:06.340 --> 00:39:11.375 align:middle line:84% And "yellow," "dim," "gray," "clear," and "brown" end in 00:39:11.375 --> 00:39:12.085 align:middle line:90% voiced sounds. 00:39:12.085 --> 00:39:14.600 align:middle line:90% 00:39:14.600 --> 00:39:18.977 align:middle line:84% What's the cover term for stops and fricatives? 00:39:18.977 --> 00:39:19.810 align:middle line:90% STUDENT: Obstruents. 00:39:19.810 --> 00:39:21.490 align:middle line:90% NORVIN RICHARDS: Obstruents. 00:39:21.490 --> 00:39:26.260 align:middle line:84% So "-en" can be added to words ending in obstruents. 00:39:26.260 --> 00:39:28.990 align:middle line:84% And it can't be added to adjectives 00:39:28.990 --> 00:39:30.730 align:middle line:90% ending in sonorants, yeah? 00:39:30.730 --> 00:39:34.290 align:middle line:90% 00:39:34.290 --> 00:39:38.280 align:middle line:84% Another place where it's useful to distinguish sonorants 00:39:38.280 --> 00:39:40.950 align:middle line:90% from obstruents-- 00:39:40.950 --> 00:39:43.740 align:middle line:84% Polish plurals-- I think this is the last time I will say 00:39:43.740 --> 00:39:45.360 align:middle line:90% anything about Polish plurals. 00:39:45.360 --> 00:39:48.910 align:middle line:84% I can't promise because it's a very useful example. 00:39:48.910 --> 00:39:51.660 align:middle line:84% But if you're getting tired of Polish plurals, I apologize. 00:39:51.660 --> 00:39:53.160 align:middle line:84% I think we're just about to the end. 00:39:53.160 --> 00:39:55.950 align:middle line:84% On the other hand, if you've become fond of Polish plurals, 00:39:55.950 --> 00:39:59.190 align:middle line:84% please gaze at this slide for a while. 00:39:59.190 --> 00:40:02.050 align:middle line:90% This is your last big chance. 00:40:02.050 --> 00:40:04.470 align:middle line:84% What we said about Polish plurals was 00:40:04.470 --> 00:40:06.480 align:middle line:84% that it's useful to say that there 00:40:06.480 --> 00:40:09.870 align:middle line:84% are words that end underlyingly in sounds like "g" and "b" 00:40:09.870 --> 00:40:10.620 align:middle line:90% and "z." 00:40:10.620 --> 00:40:12.960 align:middle line:84% And those final sounds, "g" and "b" 00:40:12.960 --> 00:40:15.300 align:middle line:84% and "z," devoice when they're at the ends of words. 00:40:15.300 --> 00:40:18.210 align:middle line:84% So underlying "wuk," the word for "lye," 00:40:18.210 --> 00:40:21.670 align:middle line:84% becomes "wuk" if you don't add anything after it. 00:40:21.670 --> 00:40:23.070 align:middle line:84% And same with the word for "club" 00:40:23.070 --> 00:40:24.630 align:middle line:90% and the word for "rubble." 00:40:24.630 --> 00:40:26.580 align:middle line:84% And I have been incautiously saying things 00:40:26.580 --> 00:40:33.130 align:middle line:84% like, "Polish devoices its final consonants," which was a lie. 00:40:33.130 --> 00:40:36.000 align:middle line:84% I think I told you just now that I try not to lie 00:40:36.000 --> 00:40:38.970 align:middle line:90% to you any more than I have to. 00:40:38.970 --> 00:40:41.730 align:middle line:84% It doesn't devoice its final consonants, right? 00:40:41.730 --> 00:40:44.910 align:middle line:84% Because Polish has words like "dom," 00:40:44.910 --> 00:40:46.440 align:middle line:90% which is the word for house. 00:40:46.440 --> 00:40:48.576 align:middle line:84% And that ends in a voiced sound, right? 00:40:48.576 --> 00:40:50.826 align:middle line:84% "Dom" doesn't become "dom" [pronounced with voicelesss 00:40:50.826 --> 00:40:52.860 align:middle line:90% "m"], yeah? 00:40:52.860 --> 00:40:54.880 align:middle line:90% Right? 00:40:54.880 --> 00:40:57.010 align:middle line:84% So it's not the case that final consonants 00:40:57.010 --> 00:40:59.568 align:middle line:90% become voiceless in Polish. 00:40:59.568 --> 00:41:01.735 align:middle line:84% What are the things that become voiceless in Polish? 00:41:01.735 --> 00:41:04.690 align:middle line:90% 00:41:04.690 --> 00:41:05.620 align:middle line:90% Obstruents, yeah? 00:41:05.620 --> 00:41:07.150 align:middle line:90% Stops and fricatives. 00:41:07.150 --> 00:41:09.790 align:middle line:84% Those are the things that become voiceless in Polish. 00:41:09.790 --> 00:41:13.540 align:middle line:84% Relatedly, Polish is like English in having voiced 00:41:13.540 --> 00:41:17.170 align:middle line:84% and voiceless obstruents, but not having 00:41:17.170 --> 00:41:18.850 align:middle line:90% voiced and voiceless sonorants. 00:41:18.850 --> 00:41:23.170 align:middle line:84% So Polish-- the "m," in a sense, doesn't have anywhere to go. 00:41:23.170 --> 00:41:24.580 align:middle line:90% Polish doesn't have that sound. 00:41:24.580 --> 00:41:27.970 align:middle line:90% 00:41:27.970 --> 00:41:29.270 align:middle line:90% Common situation. 00:41:29.270 --> 00:41:32.590 align:middle line:84% So it's really final obstruents that are becoming voiceless. 00:41:32.590 --> 00:41:36.110 align:middle line:90% 00:41:36.110 --> 00:41:37.460 align:middle line:90% OK, all right. 00:41:37.460 --> 00:41:41.030 align:middle line:84% So that's a reason to want to care 00:41:41.030 --> 00:41:44.235 align:middle line:84% about the distinction between sonorants and obstruents, 00:41:44.235 --> 00:41:45.860 align:middle line:84% not just because you want to understand 00:41:45.860 --> 00:41:48.830 align:middle line:84% why it's "whiten" and "darken" and "lessen," 00:41:48.830 --> 00:41:51.525 align:middle line:90% but not "yellowen." 00:41:51.525 --> 00:41:53.150 align:middle line:84% It seems to be a distinction that cares 00:41:53.150 --> 00:41:54.630 align:middle line:90% about sonorants and obstruents. 00:41:54.630 --> 00:41:56.090 align:middle line:90% Here's another one. 00:41:56.090 --> 00:41:58.760 align:middle line:84% Polish, which as I've said now several times, 00:41:58.760 --> 00:42:00.800 align:middle line:84% and I think this might be the last time-- 00:42:00.800 --> 00:42:03.558 align:middle line:84% Polish is doing something that is very common in languages 00:42:03.558 --> 00:42:04.100 align:middle line:90% of the world. 00:42:04.100 --> 00:42:07.850 align:middle line:90% It's devoicing final obstruents. 00:42:07.850 --> 00:42:09.650 align:middle line:84% That's a thing that happens a lot. 00:42:09.650 --> 00:42:12.980 align:middle line:90% 00:42:12.980 --> 00:42:13.480 align:middle line:90% All right. 00:42:13.480 --> 00:42:17.740 align:middle line:84% So phonology so far-- we've been doing phonology 00:42:17.740 --> 00:42:20.050 align:middle line:90% for a couple of days now. 00:42:20.050 --> 00:42:25.330 align:middle line:84% It's useful to think of sounds as undergoing changes sometimes 00:42:25.330 --> 00:42:29.030 align:middle line:84% driven by sounds coming in contact with each other. 00:42:29.030 --> 00:42:32.320 align:middle line:84% And we can describe these changes in terms of rules. 00:42:32.320 --> 00:42:35.740 align:middle line:84% There's the formalism for rules that I've been showing you. 00:42:35.740 --> 00:42:39.880 align:middle line:84% So we'll say sound A changes into sound B. 00:42:39.880 --> 00:42:43.040 align:middle line:84% The slash introduces you to the environment. 00:42:43.040 --> 00:42:45.598 align:middle line:84% This is a description of what's happening in Polish, 00:42:45.598 --> 00:42:47.140 align:middle line:84% and Turkish, and German, and Russian, 00:42:47.140 --> 00:42:48.640 align:middle line:90% and lots of other languages. 00:42:48.640 --> 00:42:51.760 align:middle line:84% Final obstruents-- so if you are an obstruent that 00:42:51.760 --> 00:42:53.560 align:middle line:84% is something which is not a sonorant, 00:42:53.560 --> 00:42:56.410 align:middle line:84% you become something which is not voiced. 00:42:56.410 --> 00:42:58.840 align:middle line:84% If you are at the end of a word-- 00:42:58.840 --> 00:43:01.300 align:middle line:84% that environment description is the way 00:43:01.300 --> 00:43:02.650 align:middle line:90% to write at the end of a word. 00:43:02.650 --> 00:43:05.950 align:middle line:84% That hash symbol is the symbol for a boundary between words. 00:43:05.950 --> 00:43:09.200 align:middle line:84% And the underscore is where the change takes place. 00:43:09.200 --> 00:43:13.270 align:middle line:84% So this rule says this change happens 00:43:13.270 --> 00:43:16.660 align:middle line:84% when the sound is before a word boundary, 00:43:16.660 --> 00:43:19.750 align:middle line:84% or to put it another way, when it is at the end of a word. 00:43:19.750 --> 00:43:21.270 align:middle line:90% Does that make sense? 00:43:21.270 --> 00:43:22.735 align:middle line:90% This is a way of writing rules. 00:43:22.735 --> 00:43:26.530 align:middle line:90% 00:43:26.530 --> 00:43:29.050 align:middle line:84% These rules are part of our knowledge of a language. 00:43:29.050 --> 00:43:30.480 align:middle line:90% So we've talked about this. 00:43:30.480 --> 00:43:34.260 align:middle line:90% 00:43:34.260 --> 00:43:39.180 align:middle line:84% As adults, anybody who is a native speaker of English 00:43:39.180 --> 00:43:42.240 align:middle line:84% knows how to make the plural of "wug," 00:43:42.240 --> 00:43:45.060 align:middle line:84% or how to make the plural of "Bach," even. 00:43:45.060 --> 00:43:48.540 align:middle line:84% So how to make plurals of words you've 00:43:48.540 --> 00:43:51.540 align:middle line:84% never heard before and even words that end in sounds 00:43:51.540 --> 00:43:54.430 align:middle line:84% that your language doesn't have, right? 00:43:54.430 --> 00:43:58.350 align:middle line:84% So if I insist on calling the composer "Bach"-- 00:43:58.350 --> 00:44:01.650 align:middle line:84% if I insist on the original Inupiaq pronunciation 00:44:01.650 --> 00:44:05.370 align:middle line:84% of the word "qaiaq," which is kayak, a single person 00:44:05.370 --> 00:44:10.810 align:middle line:84% boat, which in Inupiaq begins and ends with a uvular stop-- 00:44:10.810 --> 00:44:12.810 align:middle line:84% so if I'm trying to impress you with my ability 00:44:12.810 --> 00:44:14.550 align:middle line:84% to make uvular stops and I insist 00:44:14.550 --> 00:44:17.820 align:middle line:84% on calling a kayak a "qaiaq," then my plural for it 00:44:17.820 --> 00:44:20.370 align:middle line:90% is going to be "qaiaqs," right? 00:44:20.370 --> 00:44:21.750 align:middle line:84% So I'm going to add an "s" sound. 00:44:21.750 --> 00:44:24.030 align:middle line:84% I'm going to devoice it because that uvular stop, 00:44:24.030 --> 00:44:25.488 align:middle line:84% although it's a sound we don't have 00:44:25.488 --> 00:44:27.280 align:middle line:90% an English-- it's voiceless. 00:44:27.280 --> 00:44:31.140 align:middle line:84% So we're going to devoice after it. 00:44:31.140 --> 00:44:33.690 align:middle line:84% So these rules are part of our knowledge of a language. 00:44:33.690 --> 00:44:36.540 align:middle line:84% And we can apply them to unfamiliar problems, 00:44:36.540 --> 00:44:40.230 align:middle line:84% unfamiliar words, and even unfamiliar sounds. 00:44:40.230 --> 00:44:43.455 align:middle line:84% And I've now tried to convince you the best 00:44:43.455 --> 00:44:45.330 align:middle line:84% way to state these rules, sometimes at least, 00:44:45.330 --> 00:44:46.650 align:middle line:90% is in terms of features. 00:44:46.650 --> 00:44:49.980 align:middle line:84% That is, we don't just want to list the sounds that 00:44:49.980 --> 00:44:51.300 align:middle line:90% undergo a change. 00:44:51.300 --> 00:44:54.960 align:middle line:84% We want to have these features which are ways 00:44:54.960 --> 00:44:57.150 align:middle line:90% of capturing sets of sounds. 00:44:57.150 --> 00:45:00.840 align:middle line:84% We think that there are natural classes of sounds that 00:45:00.840 --> 00:45:02.470 align:middle line:90% undergo particular changes. 00:45:02.470 --> 00:45:04.560 align:middle line:84% So there are sound changes that apply, say, 00:45:04.560 --> 00:45:08.910 align:middle line:84% to obstruents, to sonorants, or to vowels, to voiced sounds, 00:45:08.910 --> 00:45:09.957 align:middle line:90% or to voiceless sounds. 00:45:09.957 --> 00:45:11.790 align:middle line:84% We're not just going to list the sounds that 00:45:11.790 --> 00:45:12.707 align:middle line:90% undergo sound changes. 00:45:12.707 --> 00:45:16.219 align:middle line:84% We're going to state these in terms of features, OK? 00:45:16.219 --> 00:45:20.710 align:middle line:90% 00:45:20.710 --> 00:45:23.290 align:middle line:84% And then the other thing that I've introduced you to 00:45:23.290 --> 00:45:25.840 align:middle line:84% is the idea that at least sometimes, it's 00:45:25.840 --> 00:45:29.740 align:middle line:84% fruitful to model these kinds of sound changes 00:45:29.740 --> 00:45:32.500 align:middle line:90% as involving ranked constraints. 00:45:32.500 --> 00:45:35.770 align:middle line:84% This was the optimality theory tableaux-- 00:45:35.770 --> 00:45:39.370 align:middle line:84% that is, to imagine that languages have a list of things 00:45:39.370 --> 00:45:41.000 align:middle line:90% that they are trying to achieve. 00:45:41.000 --> 00:45:44.770 align:middle line:84% And it's a list that's ranked in order of priority. 00:45:44.770 --> 00:45:47.650 align:middle line:84% And you get to observe the interplay 00:45:47.650 --> 00:45:49.420 align:middle line:84% between these different priorities 00:45:49.420 --> 00:45:51.610 align:middle line:84% in certain kinds of complicated cases. 00:45:51.610 --> 00:45:55.550 align:middle line:84% The Japanese accent case was one of the ones that we looked at. 00:45:55.550 --> 00:45:58.280 align:middle line:90% OK? 00:45:58.280 --> 00:46:00.478 align:middle line:84% And I think that's it for this review slide. 00:46:00.478 --> 00:46:01.520 align:middle line:90% Was there something else? 00:46:01.520 --> 00:46:02.810 align:middle line:90% Oh, OK. 00:46:02.810 --> 00:46:04.970 align:middle line:84% Is everyone happy with all of this? 00:46:04.970 --> 00:46:06.650 align:middle line:84% Is there any of this that people are looking at and going, 00:46:06.650 --> 00:46:07.070 align:middle line:90% no, wait. 00:46:07.070 --> 00:46:08.320 align:middle line:90% How did-- how did we see that? 00:46:08.320 --> 00:46:10.700 align:middle line:84% Or what does that mean, or anything like that? 00:46:10.700 --> 00:46:12.710 align:middle line:84% Just an attempt to make sure that we're all 00:46:12.710 --> 00:46:14.240 align:middle line:90% n the same ballpark here. 00:46:14.240 --> 00:46:15.042 align:middle line:90% Yeah? 00:46:15.042 --> 00:46:16.280 align:middle line:84% STUDENT: Could you clarify what "feature" means? 00:46:16.280 --> 00:46:17.113 align:middle line:90% NORVIN RICHARDS: Oh. 00:46:17.113 --> 00:46:20.660 align:middle line:84% All I mean by "feature" are things like voiced, 00:46:20.660 --> 00:46:23.950 align:middle line:84% or sonorant, or those kinds of things. 00:46:23.950 --> 00:46:25.370 align:middle line:90% Yeah, good question. 00:46:25.370 --> 00:46:27.170 align:middle line:84% And a standard way to write rules 00:46:27.170 --> 00:46:29.720 align:middle line:84% involves putting features in brackets like that. 00:46:29.720 --> 00:46:31.430 align:middle line:84% So the fancy way to say something 00:46:31.430 --> 00:46:35.450 align:middle line:84% is voiceless is to say that its value for the feature voice 00:46:35.450 --> 00:46:36.530 align:middle line:90% is minus. 00:46:36.530 --> 00:46:38.480 align:middle line:90% So it is not voiced, yeah. 00:46:38.480 --> 00:46:41.090 align:middle line:90% That's all that means. 00:46:41.090 --> 00:46:42.570 align:middle line:90% Yeah? 00:46:42.570 --> 00:46:43.350 align:middle line:90% OK. 00:46:43.350 --> 00:46:46.200 align:middle line:84% When we ask you to do phonology, we 00:46:46.200 --> 00:46:48.840 align:middle line:84% will tell you whether we want you-- 00:46:48.840 --> 00:46:51.150 align:middle line:84% to what extent we want you to be formal about how 00:46:51.150 --> 00:46:52.720 align:middle line:90% you write your rules. 00:46:52.720 --> 00:46:56.220 align:middle line:84% So I think the problem set that I just sent out 00:46:56.220 --> 00:47:00.270 align:middle line:84% asks you to look around for possible rules in-- 00:47:00.270 --> 00:47:02.020 align:middle line:90% what is it? 00:47:02.020 --> 00:47:05.820 align:middle line:90% It's Greenlandic data, I think. 00:47:05.820 --> 00:47:08.970 align:middle line:84% You don't have to sweat this level of formality 00:47:08.970 --> 00:47:09.730 align:middle line:90% in your rules. 00:47:09.730 --> 00:47:11.350 align:middle line:84% So if your rule for that problem set-- 00:47:11.350 --> 00:47:15.280 align:middle line:84% so if you think you've figured out what's going on, 00:47:15.280 --> 00:47:17.720 align:middle line:84% you can just tell us, more or less. 00:47:17.720 --> 00:47:19.800 align:middle line:84% Don't feel like we're going to count off 00:47:19.800 --> 00:47:24.170 align:middle line:84% for not using features or not writing your rules this way. 00:47:24.170 --> 00:47:24.670 align:middle line:90% OK. 00:47:24.670 --> 00:47:28.810 align:middle line:90% 00:47:28.810 --> 00:47:30.610 align:middle line:90% All right, then. 00:47:30.610 --> 00:47:33.465 align:middle line:90% Now I told you just now-- 00:47:33.465 --> 00:47:34.840 align:middle line:84% I tried to convince you earlier-- 00:47:34.840 --> 00:47:38.170 align:middle line:90% that there are cases where-- 00:47:38.170 --> 00:47:39.760 align:middle line:84% the Japanese accent case was the one 00:47:39.760 --> 00:47:42.100 align:middle line:84% that we spent the most time on, where 00:47:42.100 --> 00:47:45.550 align:middle line:84% it's fruitful to think of the phenomenon 00:47:45.550 --> 00:47:47.860 align:middle line:84% as involving a bunch of ranked priorities, 00:47:47.860 --> 00:47:50.610 align:middle line:84% a bunch of constraints that are all pulling 00:47:50.610 --> 00:47:51.610 align:middle line:90% in different directions. 00:47:51.610 --> 00:47:53.650 align:middle line:84% And you solve that problem by deciding 00:47:53.650 --> 00:47:56.260 align:middle line:84% which is the most important one, which 00:47:56.260 --> 00:47:58.510 align:middle line:84% ones are more important than which other ones, ranking 00:47:58.510 --> 00:48:01.970 align:middle line:84% your various constraints with respect to each other. 00:48:01.970 --> 00:48:04.270 align:middle line:84% There are other cases, though, that 00:48:04.270 --> 00:48:08.260 align:middle line:84% are problematic for that way of thinking about things. 00:48:08.260 --> 00:48:11.560 align:middle line:84% And there's big literature on trying to reconcile 00:48:11.560 --> 00:48:12.820 align:middle line:90% these kinds of observations. 00:48:12.820 --> 00:48:15.400 align:middle line:84% On the one hand, the observation that there do 00:48:15.400 --> 00:48:18.490 align:middle line:84% seem to be these conspiracies where a language is doing 00:48:18.490 --> 00:48:20.240 align:middle line:84% the same thing in a variety of ways, which 00:48:20.240 --> 00:48:22.782 align:middle line:84% is the kind of thing that makes people want to think in terms 00:48:22.782 --> 00:48:24.980 align:middle line:84% of ranked constraints, and on the other hand, 00:48:24.980 --> 00:48:27.250 align:middle line:84% the kinds of cases, which we've talked about before, 00:48:27.250 --> 00:48:29.020 align:middle line:84% where it looks more like what's happening 00:48:29.020 --> 00:48:30.580 align:middle line:84% is that you have a bunch of rules. 00:48:30.580 --> 00:48:33.040 align:middle line:84% And they're applying in an order. 00:48:33.040 --> 00:48:38.710 align:middle line:84% And they don't necessarily all get obeyed. 00:48:38.710 --> 00:48:41.510 align:middle line:90% We did this on the board-- 00:48:41.510 --> 00:48:42.010 align:middle line:90% when was it? 00:48:42.010 --> 00:48:42.802 align:middle line:90% Last time, I think. 00:48:42.802 --> 00:48:45.500 align:middle line:84% But let's do it on the slide a little more carefully. 00:48:45.500 --> 00:48:49.480 align:middle line:84% So cast your minds back to when we were talking about Lardil. 00:48:49.480 --> 00:48:51.790 align:middle line:84% I tried to convince you-- we worked out an analysis 00:48:51.790 --> 00:48:57.760 align:middle line:84% of Lardil in which, again, bare nominative singular-- 00:48:57.760 --> 00:48:58.990 align:middle line:90% Lardil doesn't mark number. 00:48:58.990 --> 00:49:02.920 align:middle line:84% But bare nominative forms of nouns 00:49:02.920 --> 00:49:04.970 align:middle line:84% undergo various kinds of sound changes. 00:49:04.970 --> 00:49:08.200 align:middle line:84% So Lardil has words that underlyingly 00:49:08.200 --> 00:49:10.090 align:middle line:90% end in a "k," for example. 00:49:10.090 --> 00:49:12.910 align:middle line:84% And there was a general rule in Lardil 00:49:12.910 --> 00:49:15.550 align:middle line:84% that said if you have a word that ends 00:49:15.550 --> 00:49:18.370 align:middle line:90% in a "k," get rid of the "k." 00:49:18.370 --> 00:49:21.400 align:middle line:84% So I'll start writing these down. 00:49:21.400 --> 00:49:24.870 align:middle line:84% Some Lardil rules-- one of them was eliminate-- 00:49:24.870 --> 00:49:26.530 align:middle line:84% and I won't try to write them formally. 00:49:26.530 --> 00:49:29.890 align:middle line:90% Eliminate final "k." 00:49:29.890 --> 00:49:32.960 align:middle line:84% So the word that we think is underlyingly 00:49:32.960 --> 00:49:35.890 align:middle line:84% "wangalk," the word for boomerang, 00:49:35.890 --> 00:49:37.700 align:middle line:84% in its nominative form is "wangal." 00:49:37.700 --> 00:49:39.617 align:middle line:84% And the reason we think that it's underlyingly 00:49:39.617 --> 00:49:41.860 align:middle line:84% "wangal" is that if you add any suffixes to it, 00:49:41.860 --> 00:49:43.270 align:middle line:90% you get to see that "k." 00:49:43.270 --> 00:49:47.380 align:middle line:84% So the accusative, I think I told you, was [LARDIL].. 00:49:47.380 --> 00:49:51.910 align:middle line:84% So the "k" shows up, and in other forms as well. 00:49:51.910 --> 00:49:54.790 align:middle line:84% We thought there was another rule that said 00:49:54.790 --> 00:50:00.460 align:middle line:90% a final "u" becomes "a," yeah? 00:50:00.460 --> 00:50:03.430 align:middle line:84% And that was why "kandu," which is 00:50:03.430 --> 00:50:05.950 align:middle line:84% the word for "blood" underlyingly, 00:50:05.950 --> 00:50:08.530 align:middle line:84% becomes "kanda" in the nominative. 00:50:08.530 --> 00:50:10.720 align:middle line:84% And in the accusative, it's [LARDIL].. 00:50:10.720 --> 00:50:13.780 align:middle line:84% That's why we think it starts out as "kandu." 00:50:13.780 --> 00:50:15.602 align:middle line:90% So we have these two rules. 00:50:15.602 --> 00:50:17.060 align:middle line:84% And what we said when we were doing 00:50:17.060 --> 00:50:20.440 align:middle line:84% Lardil before was something like we 00:50:20.440 --> 00:50:24.520 align:middle line:84% might want these rules to apply in a particular order. 00:50:24.520 --> 00:50:26.650 align:middle line:84% In fact, I see that I have written them 00:50:26.650 --> 00:50:28.750 align:middle line:90% in the wrong order. 00:50:28.750 --> 00:50:32.770 align:middle line:84% If you were to take the word "ngaluk," which 00:50:32.770 --> 00:50:35.050 align:middle line:84% is the underlying form for "story," 00:50:35.050 --> 00:50:38.510 align:middle line:84% and apply this rule and then this rule, 00:50:38.510 --> 00:50:40.660 align:middle line:84% well then, you would eliminate the final "k." 00:50:40.660 --> 00:50:42.070 align:middle line:90% That would get you ngalu. 00:50:42.070 --> 00:50:44.080 align:middle line:84% And then you would change the final "u" to "a." 00:50:44.080 --> 00:50:46.090 align:middle line:84% And then you would get an "ngala." 00:50:46.090 --> 00:50:49.000 align:middle line:84% But that's not, in fact, the nominative form for "story." 00:50:49.000 --> 00:50:52.060 align:middle line:84% The nominative form for story is ngalu, yeah? 00:50:52.060 --> 00:50:55.640 align:middle line:84% And so what I said at the time was one way of thinking 00:50:55.640 --> 00:50:57.640 align:middle line:84% about this, a popular way of thinking about this 00:50:57.640 --> 00:50:59.560 align:middle line:84% is, to think there are these rules. 00:50:59.560 --> 00:51:02.470 align:middle line:90% And it's like an assembly line. 00:51:02.470 --> 00:51:05.380 align:middle line:84% The first station is the one where you do this, 00:51:05.380 --> 00:51:07.150 align:middle line:90% where you get rid of final "u." 00:51:07.150 --> 00:51:10.150 align:middle line:84% And then after you're done with that, you get handed off 00:51:10.150 --> 00:51:12.760 align:middle line:84% to another place in the assembly line 00:51:12.760 --> 00:51:15.340 align:middle line:90% where they eliminate final "k." 00:51:15.340 --> 00:51:18.370 align:middle line:84% And what's happened with that "ngaluk," underlying "ngaluk," 00:51:18.370 --> 00:51:20.560 align:middle line:84% is that it got handed to the first station first. 00:51:20.560 --> 00:51:22.690 align:middle line:84% First station said, well, this word, "ngaluk"-- 00:51:22.690 --> 00:51:23.920 align:middle line:90% it doesn't end in "u." 00:51:23.920 --> 00:51:25.390 align:middle line:90% I'm not going to do anything. 00:51:25.390 --> 00:51:27.742 align:middle line:84% And then it gets handed to the second station. 00:51:27.742 --> 00:51:29.200 align:middle line:84% The second station says, oh, I need 00:51:29.200 --> 00:51:30.367 align:middle line:90% to get rid of the final "k." 00:51:30.367 --> 00:51:32.198 align:middle line:84% It'll be a "ngalu," by which point 00:51:32.198 --> 00:51:34.490 align:middle line:84% the people in the first station might have been saying, 00:51:34.490 --> 00:51:34.990 align:middle line:90% wait, wait. 00:51:34.990 --> 00:51:35.930 align:middle line:90% Let us have that back. 00:51:35.930 --> 00:51:37.900 align:middle line:90% But it's too late, was the idea. 00:51:37.900 --> 00:51:40.390 align:middle line:84% Yeah, so these are rules that apply in a particular order. 00:51:40.390 --> 00:51:42.820 align:middle line:84% That's the way we were talking about Lardil. 00:51:42.820 --> 00:51:45.160 align:middle line:84% Does that all sound vaguely familiar? 00:51:45.160 --> 00:51:47.785 align:middle line:84% There are these procedures, and the procedures happen in order. 00:51:47.785 --> 00:51:50.530 align:middle line:90% 00:51:50.530 --> 00:51:52.740 align:middle line:84% But if we imagine, thinking about things 00:51:52.740 --> 00:51:55.080 align:middle line:84% in an optimality theoretic kind of way, 00:51:55.080 --> 00:51:59.670 align:middle line:84% well, it's not hard to come up with constraints that say this. 00:51:59.670 --> 00:52:04.980 align:middle line:84% Star, final-- I have to write it in capital letters 00:52:04.980 --> 00:52:06.990 align:middle line:84% because that's what you do in optimality theory. 00:52:06.990 --> 00:52:08.550 align:middle line:90% No final "k." 00:52:08.550 --> 00:52:13.170 align:middle line:90% No final "u," yeah? 00:52:13.170 --> 00:52:15.550 align:middle line:84% And those constraints would look at "ngaluk"-- 00:52:15.550 --> 00:52:17.620 align:middle line:84% or sorry, "wangalk," and they would say, 00:52:17.620 --> 00:52:18.540 align:middle line:90% no, we can't have the final "k." 00:52:18.540 --> 00:52:19.290 align:middle line:90% It'll be one "wangal." 00:52:19.290 --> 00:52:20.582 align:middle line:90% And they would look at "kandu." 00:52:20.582 --> 00:52:23.220 align:middle line:84% And they would say, no, can't have the final "u." 00:52:23.220 --> 00:52:24.780 align:middle line:90% It's got to be "kanda." 00:52:24.780 --> 00:52:26.280 align:middle line:84% It might have other constraints that 00:52:26.280 --> 00:52:29.520 align:middle line:84% are responsible for helping you decide the best thing to do 00:52:29.520 --> 00:52:32.700 align:middle line:84% with "kandu" is to change it to "kanda" instead of "kandi," 00:52:32.700 --> 00:52:34.897 align:middle line:84% or "kand," or any of a variety of other things 00:52:34.897 --> 00:52:35.730 align:middle line:90% you could have done. 00:52:35.730 --> 00:52:37.380 align:middle line:84% So it's not all that hard to have 00:52:37.380 --> 00:52:39.960 align:middle line:84% constraints that say there's something wrong with "wangalk." 00:52:39.960 --> 00:52:43.640 align:middle line:84% There's something wrong with "kandu." 00:52:43.640 --> 00:52:47.400 align:middle line:84% But it's not easy to figure out what to do next. 00:52:47.400 --> 00:52:50.900 align:middle line:84% So if you have those conditions that say Lardil cannot stand 00:52:50.900 --> 00:52:55.400 align:middle line:84% words that end in "k." or words that end in "u," 00:52:55.400 --> 00:52:58.310 align:middle line:84% that's the kind of thing we were saying about Japanese accent. 00:52:58.310 --> 00:53:02.360 align:middle line:84% Japanese doesn't like accent on the last syllable. 00:53:02.360 --> 00:53:05.200 align:middle line:84% So avoid it if you're making a compound. 00:53:05.200 --> 00:53:07.900 align:middle line:84% Those generalizations hold for "boomerang" and for "blood." 00:53:07.900 --> 00:53:09.550 align:middle line:90% But they don't hold for "story." 00:53:09.550 --> 00:53:12.850 align:middle line:84% The one about not having a word ending in "u" doesn't hold. 00:53:12.850 --> 00:53:13.983 align:middle line:90% Yes? 00:53:13.983 --> 00:53:16.150 align:middle line:84% STUDENT: You talked about no final "k," no final "u" 00:53:16.150 --> 00:53:17.813 align:middle line:84% as being part of optimality theory. 00:53:17.813 --> 00:53:18.730 align:middle line:90% NORVIN RICHARDS: Yeah. 00:53:18.730 --> 00:53:21.340 align:middle line:90% 00:53:21.340 --> 00:53:26.030 align:middle line:84% STUDENT: Is it possible to satisfy both 00:53:26.030 --> 00:53:28.978 align:middle line:90% if you were to do "ngala"? 00:53:28.978 --> 00:53:31.070 align:middle line:90% NORVIN RICHARDS: Mm-hmm. 00:53:31.070 --> 00:53:36.692 align:middle line:84% STUDENT: Does optimality theory support each of the rules? 00:53:36.692 --> 00:53:38.150 align:middle line:84% NORVIN RICHARDS: So that's just it. 00:53:38.150 --> 00:53:38.900 align:middle line:90% It doesn't, right? 00:53:38.900 --> 00:53:41.005 align:middle line:84% So the conditions-- the sense in which 00:53:41.005 --> 00:53:42.380 align:middle line:84% it has something like an order is 00:53:42.380 --> 00:53:44.630 align:middle line:84% that the conditions are ranked with respect to each other. 00:53:44.630 --> 00:53:46.963 align:middle line:84% There's the one that's most important and the one that's 00:53:46.963 --> 00:53:49.320 align:middle line:90% next most important, and so on. 00:53:49.320 --> 00:53:52.910 align:middle line:84% But if we were to build a tableau, 00:53:52.910 --> 00:53:56.360 align:middle line:84% we would have these two conditions. 00:53:56.360 --> 00:53:58.910 align:middle line:90% Don't have final "k." 00:53:58.910 --> 00:54:04.370 align:middle line:90% Don't have final "u." 00:54:04.370 --> 00:54:07.850 align:middle line:84% And we would start off with "ngaluk." 00:54:07.850 --> 00:54:10.580 align:middle line:84% And we'd consider keeping "ngaluk." 00:54:10.580 --> 00:54:15.410 align:middle line:84% And we'd consider getting rid of the "k" at the end. 00:54:15.410 --> 00:54:18.470 align:middle line:84% And maybe we'd consider "ngala," right? 00:54:18.470 --> 00:54:23.030 align:middle line:84% And this one would get knocked out by star final "k," right? 00:54:23.030 --> 00:54:24.470 align:middle line:90% Because it would end in "k." 00:54:24.470 --> 00:54:28.520 align:middle line:84% And this one would get knocked out by star final "u," yeah? 00:54:28.520 --> 00:54:31.800 align:middle line:84% And this one would pass both of those conditions. 00:54:31.800 --> 00:54:34.550 align:middle line:84% So if we just did a tableau like this, if we said, 00:54:34.550 --> 00:54:36.710 align:middle line:84% Lardil seems to not like this and not like that, 00:54:36.710 --> 00:54:40.140 align:middle line:84% and we have evidence for that in these other cases, 00:54:40.140 --> 00:54:43.070 align:middle line:84% it seems as though "ngalu" ought to become "ngala." 00:54:43.070 --> 00:54:44.480 align:middle line:90% But that's wrong. 00:54:44.480 --> 00:54:47.640 align:middle line:84% So the fact is this is the one we want to win. 00:54:47.640 --> 00:54:52.550 align:middle line:84% And it's hard to see, once we've posited these conditions, how 00:54:52.550 --> 00:54:54.140 align:middle line:90% we can get this to win. 00:54:54.140 --> 00:54:57.230 align:middle line:84% This seems like it ought to be better once we've 00:54:57.230 --> 00:54:58.500 align:middle line:90% posited these conditions. 00:54:58.500 --> 00:55:00.667 align:middle line:84% So the puzzle-- and there's something people work on 00:55:00.667 --> 00:55:03.170 align:middle line:84% is try to figure out, why is that second one 00:55:03.170 --> 00:55:04.760 align:middle line:90% better than the third one? 00:55:04.760 --> 00:55:05.420 align:middle line:90% Yeah. 00:55:05.420 --> 00:55:07.823 align:middle line:84% STUDENT: How about we rephrase the no final "u" rule? 00:55:07.823 --> 00:55:08.740 align:middle line:90% NORVIN RICHARDS: Yeah. 00:55:08.740 --> 00:55:11.530 align:middle line:84% STUDENT: We rephrase it as no final "u" unless it's 00:55:11.530 --> 00:55:13.090 align:middle line:90% original form ends with a "k." 00:55:13.090 --> 00:55:14.090 align:middle line:90% NORVIN RICHARDS: Ah, OK. 00:55:14.090 --> 00:55:15.710 align:middle line:84% So we could have a rule that says-- 00:55:15.710 --> 00:55:16.700 align:middle line:90% that's nice. 00:55:16.700 --> 00:55:19.380 align:middle line:90% We could say, no final "u." 00:55:19.380 --> 00:55:21.950 align:middle line:84% This is another way of saying what you just said. 00:55:21.950 --> 00:55:28.955 align:middle line:84% No final "u" that was final in original form. 00:55:28.955 --> 00:55:32.080 align:middle line:90% 00:55:32.080 --> 00:55:32.710 align:middle line:90% Yep. 00:55:32.710 --> 00:55:35.830 align:middle line:84% So we need more capital letters, but it's OK. 00:55:35.830 --> 00:55:42.700 align:middle line:84% So yes, if we state it that way, then this 00:55:42.700 --> 00:55:45.290 align:middle line:90% passes that version of it. 00:55:45.290 --> 00:55:47.650 align:middle line:84% And as long as we have some other condition that says, 00:55:47.650 --> 00:55:49.540 align:middle line:84% don't change vowels just for the fun of it, 00:55:49.540 --> 00:55:53.840 align:middle line:84% don't change "u" because you want to, which we surely need, 00:55:53.840 --> 00:55:54.340 align:middle line:90% right? 00:55:54.340 --> 00:55:57.430 align:middle line:84% If you give Lardil a word that has 00:55:57.430 --> 00:56:01.540 align:middle line:84% an "u" in it that's not at the end, it's not going to change. 00:56:01.540 --> 00:56:03.790 align:middle line:90% Then yes, then this would win. 00:56:03.790 --> 00:56:04.630 align:middle line:90% That's right. 00:56:04.630 --> 00:56:07.720 align:middle line:84% That is indeed one of the kinds of things people do in order 00:56:07.720 --> 00:56:09.040 align:middle line:90% to handle this kind of problem. 00:56:09.040 --> 00:56:14.650 align:middle line:84% They make these conditions more complicated because now 00:56:14.650 --> 00:56:18.820 align:middle line:84% the conditions we've had so far are all about the form that 00:56:18.820 --> 00:56:19.880 align:middle line:90% gets pronounced. 00:56:19.880 --> 00:56:22.240 align:middle line:84% They say things like the form that gets pronounced 00:56:22.240 --> 00:56:24.310 align:middle line:84% had better not end in a "k," or the form that 00:56:24.310 --> 00:56:26.680 align:middle line:84% gets pronounced had better not end in a voiced obstruent 00:56:26.680 --> 00:56:29.502 align:middle line:84% if we're doing Polish, or it'd better not 00:56:29.502 --> 00:56:31.210 align:middle line:84% have two stridents at the end of the word 00:56:31.210 --> 00:56:32.807 align:middle line:90% in a language like English. 00:56:32.807 --> 00:56:34.390 align:middle line:84% This is a more complicated constraint. 00:56:34.390 --> 00:56:38.050 align:middle line:84% It says it had better not end in this kind of sound 00:56:38.050 --> 00:56:40.600 align:middle line:84% if it ended in this kind of sound 00:56:40.600 --> 00:56:44.050 align:middle line:90% originally, before we started. 00:56:44.050 --> 00:56:48.430 align:middle line:84% So there's an actual debate going on in optimality theory 00:56:48.430 --> 00:56:51.202 align:middle line:84% about whether we want to allow ourselves that power, allowing 00:56:51.202 --> 00:56:52.660 align:middle line:84% constraints that make reference not 00:56:52.660 --> 00:56:55.213 align:middle line:84% just to the pronounced version but to an earlier version. 00:56:55.213 --> 00:56:56.380 align:middle line:90% But you're absolutely right. 00:56:56.380 --> 00:57:00.160 align:middle line:84% That's a move people make, people argue for. 00:57:00.160 --> 00:57:01.990 align:middle line:84% I'm bringing this up just to show you 00:57:01.990 --> 00:57:04.000 align:middle line:84% there's this tension between these two ways 00:57:04.000 --> 00:57:06.070 align:middle line:90% of thinking about phonology. 00:57:06.070 --> 00:57:09.130 align:middle line:84% And you've just raised one of the ways, one of the ways 00:57:09.130 --> 00:57:12.070 align:middle line:84% that, indeed, has been taken in the literature 00:57:12.070 --> 00:57:15.230 align:middle line:84% to try to figure out how to deal with this tension. 00:57:15.230 --> 00:57:17.290 align:middle line:84% But if you thought that linguistics 00:57:17.290 --> 00:57:20.290 align:middle line:84% was finished, like we understood everything, no. 00:57:20.290 --> 00:57:22.540 align:middle line:84% Yeah, so this is one of the kinds of things 00:57:22.540 --> 00:57:24.760 align:middle line:84% that linguists still fight about-- 00:57:24.760 --> 00:57:26.680 align:middle line:84% how to deal with these kinds of tensions. 00:57:26.680 --> 00:57:28.470 align:middle line:90% Yes? 00:57:28.470 --> 00:57:31.040 align:middle line:84% STUDENT: I think we can still think of it as an assembly 00:57:31.040 --> 00:57:31.540 align:middle line:90% line. 00:57:31.540 --> 00:57:32.457 align:middle line:90% NORVIN RICHARDS: Yeah. 00:57:32.457 --> 00:57:34.870 align:middle line:84% STUDENT: If we think of it as an assembly line, 00:57:34.870 --> 00:57:38.780 align:middle line:84% and we throw at that, we can make it like a priority then. 00:57:38.780 --> 00:57:39.780 align:middle line:90% NORVIN RICHARDS: Uh-huh. 00:57:39.780 --> 00:57:42.500 align:middle line:84% STUDENT: We can make it like an assembly language. 00:57:42.500 --> 00:57:43.500 align:middle line:90% NORVIN RICHARDS: Uh-huh. 00:57:43.500 --> 00:57:44.935 align:middle line:90% STUDENT: We can form rules. 00:57:44.935 --> 00:57:45.810 align:middle line:90% NORVIN RICHARDS: Yes. 00:57:45.810 --> 00:57:47.583 align:middle line:84% STUDENT: Actually enforce the priority. 00:57:47.583 --> 00:57:48.750 align:middle line:90% NORVIN RICHARDS: Yeah, yeah. 00:57:48.750 --> 00:57:51.870 align:middle line:84% Yeah, so that is another move that people make. 00:57:51.870 --> 00:57:53.820 align:middle line:90% Sorry, this slide has gone away. 00:57:53.820 --> 00:57:57.980 align:middle line:90% 00:57:57.980 --> 00:58:02.130 align:middle line:84% That is another move that people make, where they say, 00:58:02.130 --> 00:58:04.820 align:middle line:84% well, there are various ways of interpreting what you just 00:58:04.820 --> 00:58:05.840 align:middle line:90% said. 00:58:05.840 --> 00:58:10.172 align:middle line:84% One move that people made, especially early on when people 00:58:10.172 --> 00:58:12.380 align:middle line:84% were worried about this problem in optimality theory, 00:58:12.380 --> 00:58:15.860 align:middle line:84% was to say we're going to do something 00:58:15.860 --> 00:58:18.770 align:middle line:84% in optimality theory that allows us to simulate 00:58:18.770 --> 00:58:20.510 align:middle line:90% having rules in an order. 00:58:20.510 --> 00:58:24.830 align:middle line:84% So what we'll do is we'll have a word go through the tableaux 00:58:24.830 --> 00:58:26.600 align:middle line:90% more than once. 00:58:26.600 --> 00:58:31.010 align:middle line:84% Or the conditions can be in different orders 00:58:31.010 --> 00:58:32.300 align:middle line:90% in the different tableaux. 00:58:32.300 --> 00:58:35.210 align:middle line:84% So we'll have a tableau first that 00:58:35.210 --> 00:58:39.440 align:middle line:84% says final "k" is very bad, let's say. 00:58:39.440 --> 00:58:43.280 align:middle line:84% And it doesn't have final "u" is bad, right? 00:58:43.280 --> 00:58:44.780 align:middle line:84% And then we'll have another tableau 00:58:44.780 --> 00:58:46.850 align:middle line:84% that maybe doesn't have final "k" is very bad, 00:58:46.850 --> 00:58:49.340 align:middle line:90% but has final "u" is bad, right? 00:58:49.340 --> 00:58:52.820 align:middle line:84% And that's a way to simulate ordered rules, 00:58:52.820 --> 00:58:54.530 align:middle line:84% basically, is to say we go through 00:58:54.530 --> 00:58:56.120 align:middle line:90% tableaux multiple times. 00:58:56.120 --> 00:58:58.215 align:middle line:84% And the tableaux look different from each other. 00:58:58.215 --> 00:58:59.840 align:middle line:84% They have different conditions in them, 00:58:59.840 --> 00:59:01.840 align:middle line:84% or they have the conditions in different orders. 00:59:01.840 --> 00:59:03.720 align:middle line:90% That's one way to talk about it. 00:59:03.720 --> 00:59:06.000 align:middle line:84% You might have had a more clever idea, though, 00:59:06.000 --> 00:59:08.120 align:middle line:90% which is the tableaux-- 00:59:08.120 --> 00:59:12.862 align:middle line:84% I was talking as though these rules are ranked with respect 00:59:12.862 --> 00:59:13.445 align:middle line:90% to each other. 00:59:13.445 --> 00:59:15.028 align:middle line:84% You know which ones are more important 00:59:15.028 --> 00:59:15.950 align:middle line:90% than which other ones. 00:59:15.950 --> 00:59:18.170 align:middle line:84% But optimality theory-- people are generally 00:59:18.170 --> 00:59:22.820 align:middle line:84% talking not as though those happen one at a time. 00:59:22.820 --> 00:59:25.760 align:middle line:84% It's just you're considering all these possibilities. 00:59:25.760 --> 00:59:28.887 align:middle line:84% And you decide which one is best by all of these conditions. 00:59:28.887 --> 00:59:31.220 align:middle line:84% You sound like you're thinking about something different 00:59:31.220 --> 00:59:34.880 align:middle line:84% where you would say you'd go back to the simple one 00:59:34.880 --> 00:59:36.980 align:middle line:84% and you'd say, we start with "ngaluk." 00:59:36.980 --> 00:59:39.860 align:middle line:90% 00:59:39.860 --> 00:59:43.700 align:middle line:84% And what we want is something that says something like, 00:59:43.700 --> 00:59:47.510 align:middle line:84% well, I should have put star final "u" first. 00:59:47.510 --> 00:59:49.145 align:middle line:90% I'll just reverse these. 00:59:49.145 --> 00:59:54.160 align:middle line:90% 00:59:54.160 --> 00:59:57.010 align:middle line:84% We have all of our candidates here. 00:59:57.010 --> 00:59:59.930 align:middle line:84% And we first consider a star final "u." 00:59:59.930 --> 01:00:00.430 align:middle line:90% Uh-oh. 01:00:00.430 --> 01:00:02.200 align:middle line:84% But if we consider a star final "u," 01:00:02.200 --> 01:00:03.700 align:middle line:84% star final "u" will look at "ngalu," 01:00:03.700 --> 01:00:06.070 align:middle line:90% and it will knock it out, right? 01:00:06.070 --> 01:00:08.320 align:middle line:90% So that will be bad right there. 01:00:08.320 --> 01:00:11.655 align:middle line:84% And we won't ever get it back if we do that. 01:00:11.655 --> 01:00:13.030 align:middle line:84% So what we want is something more 01:00:13.030 --> 01:00:15.730 align:middle line:84% like using the tableaux to do something 01:00:15.730 --> 01:00:19.240 align:middle line:84% like ordered rules, something that says 01:00:19.240 --> 01:00:20.440 align:middle line:90% we look at this first thing. 01:00:20.440 --> 01:00:25.218 align:middle line:84% And we are now basically doing rules in an order. 01:00:25.218 --> 01:00:27.260 align:middle line:84% We look at this first thing and we ask ourselves, 01:00:27.260 --> 01:00:27.927 align:middle line:90% what's "ngaluk"? 01:00:27.927 --> 01:00:31.370 align:middle line:84% Well, it doesn't violate this first rule. 01:00:31.370 --> 01:00:32.950 align:middle line:90% And so we'll keep it. 01:00:32.950 --> 01:00:35.290 align:middle line:84% And then we'll consider the next rule. 01:00:35.290 --> 01:00:36.470 align:middle line:90% It does violate that rule. 01:00:36.470 --> 01:00:37.428 align:middle line:90% And so we'll change it. 01:00:37.428 --> 01:00:38.980 align:middle line:84% We'll make a minimal change to it. 01:00:38.980 --> 01:00:41.780 align:middle line:84% This is a way of converting tableaux into ordered rules, 01:00:41.780 --> 01:00:42.280 align:middle line:90% I think. 01:00:42.280 --> 01:00:44.590 align:middle line:90% That's what we're now doing. 01:00:44.590 --> 01:00:47.415 align:middle line:90% And it could be. 01:00:47.415 --> 01:00:49.540 align:middle line:84% These are all things to think about-- open research 01:00:49.540 --> 01:00:51.320 align:middle line:90% questions. 01:00:51.320 --> 01:00:52.981 align:middle line:90% Cool. 01:00:52.981 --> 01:00:55.180 align:middle line:84% OK, so the goal here was to get all of you 01:00:55.180 --> 01:00:57.800 align:middle line:84% to try to do better than linguistics has so far done. 01:00:57.800 --> 01:00:59.290 align:middle line:84% So right now we have these two ways 01:00:59.290 --> 01:01:00.520 align:middle line:90% of thinking about phonology. 01:01:00.520 --> 01:01:02.830 align:middle line:84% And there are in some kind of conflict with each other. 01:01:02.830 --> 01:01:04.788 align:middle line:84% And there's active work on trying to figure out 01:01:04.788 --> 01:01:05.980 align:middle line:90% how to get them to mesh. 01:01:05.980 --> 01:01:10.740 align:middle line:90% 01:01:10.740 --> 01:01:12.090 align:middle line:90% OK. 01:01:12.090 --> 01:01:13.590 align:middle line:90% Cool. 01:01:13.590 --> 01:01:14.160 align:middle line:90% Oops. 01:01:14.160 --> 01:01:16.320 align:middle line:90% Where are we? 01:01:16.320 --> 01:01:17.415 align:middle line:90% Here we are. 01:01:17.415 --> 01:01:19.450 align:middle line:84% Sonorants and obstruents, we did that. 01:01:19.450 --> 01:01:20.683 align:middle line:90% We did that. 01:01:20.683 --> 01:01:24.330 align:middle line:90% OK, so let me introduce you-- 01:01:24.330 --> 01:01:25.950 align:middle line:84% we're running-- yeah, yeah, yeah. 01:01:25.950 --> 01:01:32.100 align:middle line:84% Let me introduce you to another class of sounds. 01:01:32.100 --> 01:01:36.720 align:middle line:84% Again, the point is to show you ways in which it's 01:01:36.720 --> 01:01:38.460 align:middle line:90% useful to categorize sounds. 01:01:38.460 --> 01:01:40.290 align:middle line:84% And the way to try to convince you of that 01:01:40.290 --> 01:01:44.430 align:middle line:84% is to show you phenomena that care about certain classes 01:01:44.430 --> 01:01:46.410 align:middle line:90% of sounds and not others. 01:01:46.410 --> 01:01:49.200 align:middle line:84% Here are a bunch of Arabic nouns together 01:01:49.200 --> 01:01:52.980 align:middle line:84% with their definite article, which you can see is "al." 01:01:52.980 --> 01:01:55.650 align:middle line:84% English has borrowed a bunch of Arabic words 01:01:55.650 --> 01:01:56.820 align:middle line:90% that have kept this "al." 01:01:56.820 --> 01:01:59.880 align:middle line:84% So we have words like "algebra," for example, or "algorithm," 01:01:59.880 --> 01:02:04.530 align:middle line:84% where the "al-" at the end is this word for "the." 01:02:04.530 --> 01:02:06.750 align:middle line:84% So here are some Arabic nouns with the word for "the" 01:02:06.750 --> 01:02:08.370 align:middle line:90% in front of them. 01:02:08.370 --> 01:02:12.990 align:middle line:84% But there's this other class of nouns, where the word for "the" 01:02:12.990 --> 01:02:13.720 align:middle line:90% changes. 01:02:13.720 --> 01:02:15.930 align:middle line:90% It's no longer "al." 01:02:15.930 --> 01:02:18.360 align:middle line:84% What it gets-- it still starts with "a," 01:02:18.360 --> 01:02:22.260 align:middle line:84% but the consonant in it becomes a copy of the consonant that's 01:02:22.260 --> 01:02:23.680 align:middle line:90% after it. 01:02:23.680 --> 01:02:26.790 align:middle line:84% So the word for sun, which is [ARABIC].. 01:02:26.790 --> 01:02:28.380 align:middle line:90% The sun is [ARABIC]. 01:02:28.380 --> 01:02:30.288 align:middle line:90% It's not "al-" [ARABIC]. 01:02:30.288 --> 01:02:31.830 align:middle line:84% Here I am pretending to speak Arabic. 01:02:31.830 --> 01:02:35.280 align:middle line:84% Is there anyone here who actually speaks Arabic? 01:02:35.280 --> 01:02:35.780 align:middle line:90% OK, cool. 01:02:35.780 --> 01:02:38.624 align:middle line:90% I'll just keep pretending, then. 01:02:38.624 --> 01:02:39.930 align:middle line:90% OK. 01:02:39.930 --> 01:02:41.180 align:middle line:90% This is a classic observation. 01:02:41.180 --> 01:02:45.110 align:middle line:84% Arabic linguists noticed this about their language. 01:02:45.110 --> 01:02:48.533 align:middle line:84% They name and noticed the relevant condition 01:02:48.533 --> 01:02:50.450 align:middle line:84% is that there are certain kinds of sounds that 01:02:50.450 --> 01:02:53.960 align:middle line:84% get the phenomenon on the right, where the definite article ends 01:02:53.960 --> 01:02:58.220 align:middle line:84% in a copy of the first consonant of the following noun, 01:02:58.220 --> 01:03:00.020 align:middle line:84% and other consonants that don't do that. 01:03:00.020 --> 01:03:03.170 align:middle line:84% They're called "moon letters" and "sun letters" 01:03:03.170 --> 01:03:06.140 align:middle line:84% in the Arabic tradition because well, 01:03:06.140 --> 01:03:08.745 align:middle line:84% "moon" is one of the words in the column on the left. 01:03:08.745 --> 01:03:11.120 align:middle line:84% And "sun" is one of the words in the column on the right. 01:03:11.120 --> 01:03:14.700 align:middle line:84% It's a mnemonic for these classes of sounds. 01:03:14.700 --> 01:03:17.870 align:middle line:84% So the moon letters are letters like "kuh," and "fuh," 01:03:17.870 --> 01:03:21.900 align:middle line:84% and "k," and "h," and glottal stop. 01:03:21.900 --> 01:03:25.740 align:middle line:84% And the sun letters are sounds like "sh," "d," "z," "n," 01:03:25.740 --> 01:03:26.470 align:middle line:90% and "th." 01:03:26.470 --> 01:03:32.820 align:middle line:90% 01:03:32.820 --> 01:03:34.230 align:middle line:90% The sun letters-- yes. 01:03:34.230 --> 01:03:38.550 align:middle line:84% The sun letters are the ones that I've got in the box there. 01:03:38.550 --> 01:03:41.010 align:middle line:84% They are what are called coronals. 01:03:41.010 --> 01:03:44.100 align:middle line:84% So coronals are the interdentals and alveolars 01:03:44.100 --> 01:03:45.970 align:middle line:90% and the postalveolars. 01:03:45.970 --> 01:03:47.500 align:middle line:90% If you think about them-- 01:03:47.500 --> 01:03:51.000 align:middle line:84% so feel free to sit there making those noises to yourself 01:03:51.000 --> 01:03:53.210 align:middle line:90% if you want to-- 01:03:53.210 --> 01:03:57.150 align:middle line:84% what all of them do is make a consonant sound that's 01:03:57.150 --> 01:03:59.740 align:middle line:84% articulated with the tip or the blade of your tongue-- 01:03:59.740 --> 01:04:01.350 align:middle line:84% so if the front of your tongue is 01:04:01.350 --> 01:04:05.830 align:middle line:84% touching some part of the top of your mouth or other places. 01:04:05.830 --> 01:04:07.380 align:middle line:84% So the interdentals-- your tongue 01:04:07.380 --> 01:04:09.810 align:middle line:84% is sticking out between your teeth, or the alveolars, 01:04:09.810 --> 01:04:12.480 align:middle line:84% or the postalveolars-- the tip of your tongue 01:04:12.480 --> 01:04:15.750 align:middle line:84% or the blade of your tongue is coming into contact almost 01:04:15.750 --> 01:04:20.225 align:middle line:84% with the top of your mouth as opposed to, say, bilabials, 01:04:20.225 --> 01:04:22.350 align:middle line:84% where your tongue can do whatever the heck it feels 01:04:22.350 --> 01:04:27.420 align:middle line:84% like, or velars, which involve your tongue, or uvulars, 01:04:27.420 --> 01:04:28.560 align:middle line:90% which involve your tongue. 01:04:28.560 --> 01:04:31.020 align:middle line:84% But it's the back of your tongue, "k," 01:04:31.020 --> 01:04:33.258 align:middle line:84% or "g," as opposed to the front of your tongue, which 01:04:33.258 --> 01:04:34.966 align:middle line:84% is what you're using for things like "t," 01:04:34.966 --> 01:04:38.790 align:middle line:84% and "d," and "f," and "th" and "sh" and "zh." 01:04:38.790 --> 01:04:39.660 align:middle line:90% Is that clear? 01:04:39.660 --> 01:04:41.340 align:middle line:90% That distinction clear? 01:04:41.340 --> 01:04:47.340 align:middle line:84% And Arabic gives this a nice illustration of this. 01:04:47.340 --> 01:04:50.430 align:middle line:84% The coronal sounds are the ones that are the sun letters. 01:04:50.430 --> 01:04:54.300 align:middle line:84% They're the ones where the last consonant 01:04:54.300 --> 01:04:56.400 align:middle line:84% of the definite article just becomes a copy 01:04:56.400 --> 01:04:58.170 align:middle line:90% of the following consonant. 01:04:58.170 --> 01:05:00.780 align:middle line:84% Notice that the last consonant of the indefinite article 01:05:00.780 --> 01:05:04.380 align:middle line:84% underlyingly is "l," which is itself coronal. 01:05:04.380 --> 01:05:08.070 align:middle line:84% So it's like the end of the definite article 01:05:08.070 --> 01:05:09.630 align:middle line:90% is going to be coronal. 01:05:09.630 --> 01:05:11.983 align:middle line:84% If nothing else happens, it's going to be "l," 01:05:11.983 --> 01:05:13.650 align:middle line:84% But if there's a coronal right after it, 01:05:13.650 --> 01:05:15.930 align:middle line:84% it just becomes a copy of the following coronal. 01:05:15.930 --> 01:05:17.730 align:middle line:90% Does that make sense? 01:05:17.730 --> 01:05:20.440 align:middle line:84% That's what's going on in all of these examples. 01:05:20.440 --> 01:05:22.560 align:middle line:84% So with "moon," and "mare," and "book," and "war," 01:05:22.560 --> 01:05:25.650 align:middle line:84% and "father," the "l" of the definite article 01:05:25.650 --> 01:05:28.290 align:middle line:84% can't become a copy of the following consonant 01:05:28.290 --> 01:05:32.370 align:middle line:84% and remain coronal, whereas in the sun letters, 01:05:32.370 --> 01:05:34.230 align:middle line:84% the ones on the right, by copying 01:05:34.230 --> 01:05:37.470 align:middle line:84% the following consonant, the "l" of the definite article 01:05:37.470 --> 01:05:40.260 align:middle line:90% is still a coronal. 01:05:40.260 --> 01:05:41.010 align:middle line:90% [ELECTRONIC CHIME] 01:05:41.010 --> 01:05:42.450 align:middle line:90% Bling. 01:05:42.450 --> 01:05:44.040 align:middle line:84% Somebody had a great idea about that. 01:05:44.040 --> 01:05:46.920 align:middle line:90% 01:05:46.920 --> 01:05:49.775 align:middle line:84% So that's what's going on in these Arabic examples. 01:05:49.775 --> 01:05:52.802 align:middle line:90% 01:05:52.802 --> 01:05:55.260 align:middle line:84% A rule that we could have for the Arabic definite article-- 01:05:55.260 --> 01:05:56.070 align:middle line:90% yes, I just said this. 01:05:56.070 --> 01:05:56.570 align:middle line:90% Thank you. 01:05:56.570 --> 01:05:58.013 align:middle line:90% "l" is also coronal. 01:05:58.013 --> 01:06:00.180 align:middle line:84% A rule we could have for the Arabic definite article 01:06:00.180 --> 01:06:02.400 align:middle line:84% would say in the definite article, 01:06:02.400 --> 01:06:06.610 align:middle line:84% an "l" becomes a copy of a following coronal consonant. 01:06:06.610 --> 01:06:10.710 align:middle line:84% So we have a variable, consonant subscript "i" 01:06:10.710 --> 01:06:13.920 align:middle line:84% which says whatever the following consonant is if it's 01:06:13.920 --> 01:06:19.200 align:middle line:84% coronal, the "l' will also get the value consonant sub i, 01:06:19.200 --> 01:06:22.510 align:middle line:84% but it'll have the same form as the following consonant. 01:06:22.510 --> 01:06:23.010 align:middle line:90% Yeah? 01:06:23.010 --> 01:06:24.750 align:middle line:84% STUDENT: I had a question about coronals. 01:06:24.750 --> 01:06:28.982 align:middle line:84% So what I noticed is the sun letters 01:06:28.982 --> 01:06:31.232 align:middle line:84% are the letters that you can maintain for a long time, 01:06:31.232 --> 01:06:35.760 align:middle line:84% like "sh," and like the "sh," like [ARABIC] in sun, 01:06:35.760 --> 01:06:38.013 align:middle line:90% for "d" in [ARABIC]. 01:06:38.013 --> 01:06:38.930 align:middle line:90% NORVIN RICHARDS: Yeah. 01:06:38.930 --> 01:06:40.580 align:middle line:84% STUDENT: You maintain them for a long time. 01:06:40.580 --> 01:06:41.497 align:middle line:90% NORVIN RICHARDS: Yeah. 01:06:41.497 --> 01:06:44.210 align:middle line:84% STUDENT: And Arabic actually has [ARABIC],, 01:06:44.210 --> 01:06:46.383 align:middle line:90% which duplicates one letter. 01:06:46.383 --> 01:06:47.300 align:middle line:90% NORVIN RICHARDS: Yeah. 01:06:47.300 --> 01:06:49.610 align:middle line:84% STUDENT: So what I'm thinking is that when 01:06:49.610 --> 01:06:51.885 align:middle line:84% you can duplicate the letter, you just duplicate it. 01:06:51.885 --> 01:06:52.300 align:middle line:90% NORVIN RICHARDS: Uh-huh. 01:06:52.300 --> 01:06:54.110 align:middle line:84% STUDENT: With [ARABIC],, and other words. 01:06:54.110 --> 01:06:55.250 align:middle line:90% NORVIN RICHARDS: Yeah. 01:06:55.250 --> 01:07:00.320 align:middle line:84% STUDENT: But for [ARABIC],, just the moon, you can't do that. 01:07:00.320 --> 01:07:01.370 align:middle line:90% NORVIN RICHARDS: Yeah. 01:07:01.370 --> 01:07:02.510 align:middle line:90% But there are-- 01:07:02.510 --> 01:07:03.290 align:middle line:84% STUDENT: Do you see what I'm saying with that? 01:07:03.290 --> 01:07:04.832 align:middle line:84% NORVIN RICHARDS: I see what you mean. 01:07:04.832 --> 01:07:05.360 align:middle line:90% But look. 01:07:05.360 --> 01:07:12.230 align:middle line:84% There are consonants like "p," and "t," and "k," which are all 01:07:12.230 --> 01:07:13.190 align:middle line:90% stops. 01:07:13.190 --> 01:07:17.450 align:middle line:84% But "t" is coronal and "p" and "k" are not, right? 01:07:17.450 --> 01:07:20.870 align:middle line:84% And so all of those are equally hard to maintain 01:07:20.870 --> 01:07:22.190 align:middle line:90% for a long time, I think. 01:07:22.190 --> 01:07:24.140 align:middle line:84% Or similarly, I think what we're hoping-- 01:07:24.140 --> 01:07:26.270 align:middle line:84% and you can stop me if this is not true-- 01:07:26.270 --> 01:07:31.160 align:middle line:84% is that if there are Arabic words that start with "m," 01:07:31.160 --> 01:07:35.210 align:middle line:84% you'll get al before an "m," yeah? 01:07:35.210 --> 01:07:37.550 align:middle line:84% But for Arabic words that start with "n," 01:07:37.550 --> 01:07:39.770 align:middle line:90% you get a copy of the "n," yeah? 01:07:39.770 --> 01:07:43.340 align:middle line:84% And again, those are all equally easy to hold on 01:07:43.340 --> 01:07:44.360 align:middle line:90% for a long time. 01:07:44.360 --> 01:07:46.220 align:middle line:84% So this is a good example of science. 01:07:46.220 --> 01:07:47.610 align:middle line:90% You had a good hypothesis. 01:07:47.610 --> 01:07:49.760 align:middle line:84% And now we're finding data, yeah. 01:07:49.760 --> 01:07:51.530 align:middle line:84% And I should have called on you to do 01:07:51.530 --> 01:07:53.390 align:middle line:90% to say these Arabic words. 01:07:53.390 --> 01:07:54.623 align:middle line:90% Yeah, Joseph? 01:07:54.623 --> 01:07:57.762 align:middle line:84% STUDENT: So you mentioned that the English language 01:07:57.762 --> 01:08:00.013 align:middle line:84% has borrowed quite a few words from Arabic. 01:08:00.013 --> 01:08:00.930 align:middle line:90% NORVIN RICHARDS: Yeah. 01:08:00.930 --> 01:08:04.980 align:middle line:84% STUDENT: And we accidentally brought the definite article 01:08:04.980 --> 01:08:05.870 align:middle line:90% with it. 01:08:05.870 --> 01:08:06.970 align:middle line:90% So the word "algebra." 01:08:06.970 --> 01:08:08.866 align:middle line:90% NORVIN RICHARDS: Yep. 01:08:08.866 --> 01:08:10.282 align:middle line:90% STUDENT: "Algebra." 01:08:10.282 --> 01:08:12.642 align:middle line:84% At least, the English pronunciation 01:08:12.642 --> 01:08:14.217 align:middle line:90% starts with coronal. 01:08:14.217 --> 01:08:15.300 align:middle line:90% NORVIN RICHARDS: Oh, yeah. 01:08:15.300 --> 01:08:16.920 align:middle line:84% But in Arabic, I think it doesn't. 01:08:16.920 --> 01:08:20.279 align:middle line:90% So maybe you know. 01:08:20.279 --> 01:08:22.740 align:middle line:84% Do you know know what "algebra" is in Arabic? 01:08:22.740 --> 01:08:24.428 align:middle line:84% STUDENT: I believe it's [ARABIC].. 01:08:24.428 --> 01:08:25.470 align:middle line:90% NORVIN RICHARDS: [ARABIC] 01:08:25.470 --> 01:08:26.178 align:middle line:90% STUDENT: [ARABIC] 01:08:26.178 --> 01:08:27.240 align:middle line:90% NORVIN RICHARDS: Oh, OK. 01:08:27.240 --> 01:08:28.229 align:middle line:90% So I got it backwards. 01:08:28.229 --> 01:08:32.100 align:middle line:84% Yeah, so we kept the "l," We didn't do the change. 01:08:32.100 --> 01:08:34.982 align:middle line:84% So it starts with a "j," starts with a coronal. 01:08:34.982 --> 01:08:36.149 align:middle line:90% But it didn't do the change. 01:08:36.149 --> 01:08:38.819 align:middle line:84% I wonder whether we borrowed it before Arabic 01:08:38.819 --> 01:08:40.720 align:middle line:90% did this sound change. 01:08:40.720 --> 01:08:42.069 align:middle line:90% I have no idea. 01:08:42.069 --> 01:08:44.540 align:middle line:84% I don't know enough about the history of Arabic. 01:08:44.540 --> 01:08:46.270 align:middle line:90% Cool. 01:08:46.270 --> 01:08:46.770 align:middle line:90% Yep. 01:08:46.770 --> 01:08:51.609 align:middle line:90% 01:08:51.609 --> 01:08:54.750 align:middle line:90% All right. 01:08:54.750 --> 01:08:57.180 align:middle line:90% So this is one rule. 01:08:57.180 --> 01:09:00.180 align:middle line:84% It's a very common rule, where one sound becomes more 01:09:00.180 --> 01:09:01.290 align:middle line:90% like another sound. 01:09:01.290 --> 01:09:03.790 align:middle line:90% It's called assimilation. 01:09:03.790 --> 01:09:06.540 align:middle line:84% And in this particular case, it's 01:09:06.540 --> 01:09:09.180 align:middle line:84% what's sometimes called total assimilation, where 01:09:09.180 --> 01:09:12.540 align:middle line:84% one sound just becomes a complete copy of another sound. 01:09:12.540 --> 01:09:15.720 align:middle line:84% We'll see other cases where one sound just 01:09:15.720 --> 01:09:20.100 align:middle line:84% changes in ways that resemble a nearby sound. 01:09:20.100 --> 01:09:21.479 align:middle line:90% Why do you do assimilation? 01:09:21.479 --> 01:09:23.470 align:middle line:84% Languages do assimilation all the time. 01:09:23.470 --> 01:09:27.090 align:middle line:84% And it's not hard to have pretheoretical notions 01:09:27.090 --> 01:09:28.350 align:middle line:90% about why, right? 01:09:28.350 --> 01:09:32.160 align:middle line:84% You're saving wear and tear on your articulators. 01:09:32.160 --> 01:09:34.920 align:middle line:84% So" making it so that your articulators-- your tongue, 01:09:34.920 --> 01:09:40.770 align:middle line:84% your lips, your velum don't have to do as much moving around. 01:09:40.770 --> 01:09:45.342 align:middle line:84% Things don't move as far or don't move at all. 01:09:45.342 --> 01:09:47.550 align:middle line:84% It probably also makes perception easier in the sense 01:09:47.550 --> 01:09:53.850 align:middle line:84% that if you're speaking Arabic and instead of an "l"-- 01:09:53.850 --> 01:09:56.490 align:middle line:90% so let's take the sun. 01:09:56.490 --> 01:09:59.190 align:middle line:90% 01:09:59.190 --> 01:10:03.380 align:middle line:84% If you know that the word for "sun" is [ARABIC],, 01:10:03.380 --> 01:10:07.730 align:middle line:84% and you've got this thing before it, 01:10:07.730 --> 01:10:09.770 align:middle line:84% basically, it's as though you only 01:10:09.770 --> 01:10:12.080 align:middle line:84% need to give your hearer enough information 01:10:12.080 --> 01:10:13.820 align:middle line:84% to help them to know that they're 01:10:13.820 --> 01:10:17.180 align:middle line:84% hearing a definite article before the word for sun. 01:10:17.180 --> 01:10:20.270 align:middle line:84% And Arabic has decided it's enough 01:10:20.270 --> 01:10:25.430 align:middle line:84% if you hear the glottal stop and the vowel, and then a coronal. 01:10:25.430 --> 01:10:27.980 align:middle line:84% Underlyingly, that coronal, maybe, is an "l." 01:10:27.980 --> 01:10:29.510 align:middle line:84% But as long as we keep it coronal, 01:10:29.510 --> 01:10:32.550 align:middle line:84% we've given the hearer enough that they get to hear it. 01:10:32.550 --> 01:10:35.720 align:middle line:84% And in fact, we give the hearer, in a sense, 01:10:35.720 --> 01:10:39.140 align:middle line:84% two shots at hearing the beginning of the word for sun. 01:10:39.140 --> 01:10:42.920 align:middle line:90% So we'll lengthen that sound. 01:10:42.920 --> 01:10:45.350 align:middle line:90% We'll pronounce it twice. 01:10:45.350 --> 01:10:49.610 align:middle line:84% So we're redistributing the signal, in a way, 01:10:49.610 --> 01:10:53.428 align:middle line:84% to the hearer to make their life arguably easier, 01:10:53.428 --> 01:10:55.970 align:middle line:84% giving up some of what makes the definite article distinctive 01:10:55.970 --> 01:10:58.550 align:middle line:84% in order to emphasize some of the beginning 01:10:58.550 --> 01:10:59.390 align:middle line:90% of the word for sun. 01:10:59.390 --> 01:11:01.340 align:middle line:90% Yeah. 01:11:01.340 --> 01:11:02.820 align:middle line:90% Questions about any of this? 01:11:02.820 --> 01:11:03.320 align:middle line:90% Yes. 01:11:03.320 --> 01:11:04.778 align:middle line:84% STUDENT: So would you say generally 01:11:04.778 --> 01:11:07.980 align:middle line:84% that the focus is just to facilitate the pronunciation 01:11:07.980 --> 01:11:09.700 align:middle line:90% of the word easier. 01:11:09.700 --> 01:11:12.340 align:middle line:84% NORVIN RICHARDS: So I was trying to float 01:11:12.340 --> 01:11:15.520 align:middle line:84% the idea that it's partly to make the pronunciation easier. 01:11:15.520 --> 01:11:17.860 align:middle line:84% It's so that your articulators are not having 01:11:17.860 --> 01:11:20.110 align:middle line:90% to move as fast or as far. 01:11:20.110 --> 01:11:24.130 align:middle line:84% But it might also make perception, in some ways, 01:11:24.130 --> 01:11:26.240 align:middle line:90% easier in a case like this. 01:11:26.240 --> 01:11:30.640 align:middle line:84% So the idea is we'll give up some of the special properties 01:11:30.640 --> 01:11:34.450 align:middle line:84% of the definite article in favor of emphasizing or holding 01:11:34.450 --> 01:11:38.590 align:middle line:84% for longer the special properties of the word for sun, 01:11:38.590 --> 01:11:40.060 align:middle line:84% or whatever, the consonant that's 01:11:40.060 --> 01:11:44.260 align:middle line:84% at the beginning of the following noun. 01:11:44.260 --> 01:11:47.440 align:middle line:84% So we'll emphasize some parts of the-- 01:11:47.440 --> 01:11:49.510 align:middle line:84% in some ways, the least predictable part of this. 01:11:49.510 --> 01:11:51.635 align:middle line:84% You presumably hear definite articles all the time. 01:11:51.635 --> 01:11:55.000 align:middle line:84% It's not so important that you get lots of clues to the fact 01:11:55.000 --> 01:11:56.440 align:middle line:90% that this is a definite article. 01:11:56.440 --> 01:11:58.000 align:middle line:90% The nouns are going to vary. 01:11:58.000 --> 01:12:02.300 align:middle line:84% You'll hear more different kinds of nouns, yeah. 01:12:02.300 --> 01:12:03.020 align:middle line:90% Yep. 01:12:03.020 --> 01:12:03.997 align:middle line:90% OK, cool. 01:12:03.997 --> 01:12:05.580 align:middle line:84% This is probably a good place to stop. 01:12:05.580 --> 01:12:06.372 align:middle line:90% So let's stop here. 01:12:06.372 --> 01:12:08.250 align:middle line:90% And have a good weekend. 01:12:08.250 --> 01:12:11.680 align:middle line:90% And I'll see you guys next week. 01:12:11.680 --> 01:12:19.000 align:middle line:90%