WEBVTT 00:00:00.000 --> 00:00:01.916 align:middle line:90% [SQUEAKING] 00:00:01.916 --> 00:00:03.832 align:middle line:90% [RUSTLING] 00:00:03.832 --> 00:00:07.185 align:middle line:90% [CLICKING] 00:00:07.185 --> 00:00:11.510 align:middle line:90% 00:00:11.510 --> 00:00:12.920 align:middle line:90% NORVIN RICHARDS: All right. 00:00:12.920 --> 00:00:17.840 align:middle line:84% So last time, we were talking about English plurals. 00:00:17.840 --> 00:00:20.270 align:middle line:84% And we went through this fairly fast 00:00:20.270 --> 00:00:23.420 align:middle line:84% so I wanted to talk about it again a little 00:00:23.420 --> 00:00:27.350 align:middle line:84% bit just to make sure we're all on the same page. 00:00:27.350 --> 00:00:33.860 align:middle line:84% The descriptive fact is that English plurals, putting aside 00:00:33.860 --> 00:00:36.860 align:middle line:84% "children," and "oxen," and "fish," and "sheep," 00:00:36.860 --> 00:00:38.540 align:middle line:84% and other things like that, there's 00:00:38.540 --> 00:00:42.380 align:middle line:84% a general plural that's used most of the time in English 00:00:42.380 --> 00:00:44.660 align:middle line:84% that we typically spell with an S. 00:00:44.660 --> 00:00:46.170 align:middle line:84% But you've heard me talk about how 00:00:46.170 --> 00:00:47.420 align:middle line:90% I feel about English spelling. 00:00:47.420 --> 00:00:52.040 align:middle line:90% 00:00:52.040 --> 00:00:55.010 align:middle line:84% It's pronounced in at least three different ways, sometimes 00:00:55.010 --> 00:00:57.470 align:middle line:84% as a "z," like in "dogs," sometimes 00:00:57.470 --> 00:01:01.250 align:middle line:84% as an "s," like in cats or giraffes, and sometimes 00:01:01.250 --> 00:01:04.280 align:middle line:84% as a schwa followed by a "z," like in "bridges," 00:01:04.280 --> 00:01:06.560 align:middle line:90% or "brushes," or "messes." 00:01:06.560 --> 00:01:10.970 align:middle line:84% And so descriptively, it's schwa "z" 00:01:10.970 --> 00:01:12.590 align:middle line:90% after what's called a strident. 00:01:12.590 --> 00:01:14.930 align:middle line:84% So the stridents are a particular class 00:01:14.930 --> 00:01:17.750 align:middle line:84% of fricatives that have a lot of high frequency noise. 00:01:17.750 --> 00:01:21.440 align:middle line:84% They are "s," and "sh," and also the 00:01:21.440 --> 00:01:25.190 align:middle line:84% affricates that end in those fricatives, the "chu," 00:01:25.190 --> 00:01:30.990 align:middle line:84% and the "juh," "tsuh" for that matter. 00:01:30.990 --> 00:01:33.080 align:middle line:90% Those are all stridents. 00:01:33.080 --> 00:01:35.945 align:middle line:84% And so you get schwa "z" after a strident, 00:01:35.945 --> 00:01:37.820 align:middle line:84% and then if you're not looking at a strident, 00:01:37.820 --> 00:01:39.050 align:middle line:90% it's just voicing. 00:01:39.050 --> 00:01:42.770 align:middle line:84% You get "s" if the sound before the plural is not strident, 00:01:42.770 --> 00:01:45.950 align:middle line:84% is voiceless, and you get "z" if it's voiced. 00:01:45.950 --> 00:01:48.240 align:middle line:84% And we had talked about all that before. 00:01:48.240 --> 00:01:50.720 align:middle line:84% And what I asked you to do last time, 00:01:50.720 --> 00:01:52.970 align:middle line:84% and I just want to go through this exercise again real 00:01:52.970 --> 00:01:56.420 align:middle line:84% quick just to make sure we're all following it. 00:01:56.420 --> 00:01:57.560 align:middle line:90% Oh, I get to take this off. 00:01:57.560 --> 00:02:00.620 align:middle line:90% I keep forgetting that. 00:02:00.620 --> 00:02:04.670 align:middle line:84% Want to go through the story again real quick. 00:02:04.670 --> 00:02:07.700 align:middle line:84% Forget for a second about the strident thing. 00:02:07.700 --> 00:02:12.110 align:middle line:84% We said there are at least two imaginable stories. 00:02:12.110 --> 00:02:14.483 align:middle line:84% We're in a room full of very imaginative people. 00:02:14.483 --> 00:02:16.400 align:middle line:84% There are lots and lots of imaginable stories, 00:02:16.400 --> 00:02:18.140 align:middle line:84% but let's concentrate on two that 00:02:18.140 --> 00:02:20.300 align:middle line:90% seem particularly attractive. 00:02:20.300 --> 00:02:24.530 align:middle line:84% One would be, the plural suffix is underlyingly an "s,", 00:02:24.530 --> 00:02:27.410 align:middle line:84% and after a voiced sound, it becomes a "z." 00:02:27.410 --> 00:02:29.180 align:middle line:84% So it starts off as an "s,", that's 00:02:29.180 --> 00:02:30.590 align:middle line:84% the one you're getting in "cats." 00:02:30.590 --> 00:02:34.610 align:middle line:84% And when you put the "s" after the "g" in "dog," 00:02:34.610 --> 00:02:36.980 align:middle line:90% it voices to a "z." 00:02:36.980 --> 00:02:39.360 align:middle line:84% And the other imaginable story would be the opposite. 00:02:39.360 --> 00:02:41.690 align:middle line:84% It would say the plural is underlyingly a "z," 00:02:41.690 --> 00:02:45.350 align:middle line:84% and after a voiceless consonant it becomes an "s." 00:02:45.350 --> 00:02:46.910 align:middle line:84% So that would say the "z" that you're 00:02:46.910 --> 00:02:51.160 align:middle line:84% getting in "dogs" is the basic plural suffix. 00:02:51.160 --> 00:02:53.560 align:middle line:84% And it's devoicing in words like "cats" 00:02:53.560 --> 00:02:56.830 align:middle line:84% because it's preceded by a voiceless consonant. 00:02:56.830 --> 00:02:59.560 align:middle line:84% And we convinced ourselves that one of these is-- 00:02:59.560 --> 00:03:01.480 align:middle line:84% or I tried to convince you, let me put it 00:03:01.480 --> 00:03:04.900 align:middle line:84% that way, that one of these is more attractive than the other, 00:03:04.900 --> 00:03:09.790 align:middle line:84% because one of these allows us to treat 00:03:09.790 --> 00:03:13.870 align:middle line:84% the sound-changing question as a general property 00:03:13.870 --> 00:03:17.380 align:middle line:84% of English phonology, rather than as something we 00:03:17.380 --> 00:03:20.450 align:middle line:84% must specifically state about this morpheme. 00:03:20.450 --> 00:03:22.870 align:middle line:84% So I'll just go through this again. 00:03:22.870 --> 00:03:27.580 align:middle line:84% On a theory where the suffix is an "s" 00:03:27.580 --> 00:03:32.470 align:middle line:84% and it voices after a voiced sound, we ask ourselves, 00:03:32.470 --> 00:03:34.810 align:middle line:84% could we make that a general fact about English? 00:03:34.810 --> 00:03:37.570 align:middle line:84% That is, is it generally true that if you 00:03:37.570 --> 00:03:40.060 align:middle line:84% have an "s" at the end of a word, 00:03:40.060 --> 00:03:43.000 align:middle line:84% you can't have a voiced sound before it? 00:03:43.000 --> 00:03:45.610 align:middle line:84% Are there any English words that end 00:03:45.610 --> 00:03:49.630 align:middle line:84% with an "s," and before the "s" there's something voiced? 00:03:49.630 --> 00:03:56.630 align:middle line:84% Can anybody come up with one, a word like that? 00:03:56.630 --> 00:03:58.550 align:middle line:84% Do people see why I'm looking for one? 00:03:58.550 --> 00:03:59.134 align:middle line:90% Yeah, Faith? 00:03:59.134 --> 00:04:00.672 align:middle line:90% AUDIENCE: What about "ribs"? 00:04:00.672 --> 00:04:02.130 align:middle line:84% NORVIN RICHARDS: So ribs, if I were 00:04:02.130 --> 00:04:05.010 align:middle line:84% going to write that in IPA, would be that. 00:04:05.010 --> 00:04:06.372 align:middle line:90% That's a "z" sound. 00:04:06.372 --> 00:04:08.580 align:middle line:84% So is there a word where there's an "z" sound that's, 00:04:08.580 --> 00:04:09.930 align:middle line:84% right before it, there's something voiced? 00:04:09.930 --> 00:04:10.185 align:middle line:90% Yes? 00:04:10.185 --> 00:04:11.140 align:middle line:90% AUDIENCE: "Tense." 00:04:11.140 --> 00:04:12.182 align:middle line:90% NORVIN RICHARDS: "Tense." 00:04:12.182 --> 00:04:14.360 align:middle line:90% Yeah, that's a good one. 00:04:14.360 --> 00:04:16.300 align:middle line:90% So that's "tense" in IPA. 00:04:16.300 --> 00:04:18.850 align:middle line:84% I guess if I'm being more specific, 00:04:18.850 --> 00:04:21.069 align:middle line:84% there's an aspirated "t" there, "tense." 00:04:21.069 --> 00:04:24.535 align:middle line:84% OK, so that's one where that could be the right theory, 00:04:24.535 --> 00:04:27.160 align:middle line:84% but it would have to be a theory that made a specific reference 00:04:27.160 --> 00:04:29.800 align:middle line:84% to this morpheme that said, special thing about the plural 00:04:29.800 --> 00:04:32.020 align:middle line:84% morpheme, it voices after a voiced sound. 00:04:32.020 --> 00:04:34.150 align:middle line:84% Because there's nothing about English phonology 00:04:34.150 --> 00:04:36.910 align:middle line:84% that prevents us from having words 00:04:36.910 --> 00:04:40.300 align:middle line:84% that end in an "s" immediately preceded by a voiced sound. 00:04:40.300 --> 00:04:42.670 align:middle line:84% Suppose we do it the other way around. 00:04:42.670 --> 00:04:46.180 align:middle line:84% If the plural is underlyingly a "z" and then 00:04:46.180 --> 00:04:49.150 align:middle line:84% we say it devoices after a voiceless consonant. 00:04:49.150 --> 00:04:50.950 align:middle line:84% Can we think of any English words 00:04:50.950 --> 00:04:52.900 align:middle line:84% that end in a "z," and before the "z," 00:04:52.900 --> 00:04:55.810 align:middle line:90% there's something voiceless? 00:04:55.810 --> 00:04:57.442 align:middle line:90% Sorry, you have another comment? 00:04:57.442 --> 00:04:59.650 align:middle line:84% AUDIENCE: Are we not thinking something like plushes, 00:04:59.650 --> 00:05:01.673 align:middle line:84% or something that's like [INAUDIBLE]?? 00:05:01.673 --> 00:05:03.340 align:middle line:84% NORVIN RICHARDS: OK, so plushes, suppose 00:05:03.340 --> 00:05:07.690 align:middle line:84% I write that down, "pluh," "shes." 00:05:07.690 --> 00:05:11.350 align:middle line:84% This is an instance of what I asked you to not think about 00:05:11.350 --> 00:05:15.160 align:middle line:84% for a while, but you're a rebel, I'm sorry. 00:05:15.160 --> 00:05:18.070 align:middle line:84% Here we've got the plural suffix and it's after a strident, 00:05:18.070 --> 00:05:20.228 align:middle line:84% so we're getting this schwa that's 00:05:20.228 --> 00:05:22.270 align:middle line:84% put in here because there's a strident before it. 00:05:22.270 --> 00:05:25.220 align:middle line:84% I started this by saying, let's put that aside for now 00:05:25.220 --> 00:05:29.720 align:middle line:84% and we'll come back to it, so good point. 00:05:29.720 --> 00:05:31.790 align:middle line:84% But actually, this isn't an example of a word 00:05:31.790 --> 00:05:33.770 align:middle line:84% that ends in a Z and before there's something voiceless 00:05:33.770 --> 00:05:34.728 align:middle line:90% because this is voiced. 00:05:34.728 --> 00:05:37.430 align:middle line:84% So we're looking for a word that ends in a Z, and before the Z, 00:05:37.430 --> 00:05:41.130 align:middle line:84% there's something voiceless, sort of the opposite of tense. 00:05:41.130 --> 00:05:43.570 align:middle line:90% Are there any words like that? 00:05:43.570 --> 00:05:45.384 align:middle line:90% I heard an "uh." 00:05:45.384 --> 00:05:46.680 align:middle line:90% AUDIENCE: "Blitz"? 00:05:46.680 --> 00:05:47.722 align:middle line:90% NORVIN RICHARDS: Sorry? 00:05:47.722 --> 00:05:49.680 align:middle line:84% AUDIENCE: I don't know, I was thinking "blitz," 00:05:49.680 --> 00:05:50.810 align:middle line:90% but that's [INAUDIBLE]. 00:05:50.810 --> 00:05:51.260 align:middle line:90% NORVIN RICHARDS: "Blitz"? 00:05:51.260 --> 00:05:51.760 align:middle line:90% OK. 00:05:51.760 --> 00:05:54.110 align:middle line:84% Well, so that's a word that when we spell it, 00:05:54.110 --> 00:05:55.780 align:middle line:90% we spell it like this. 00:05:55.780 --> 00:05:57.530 align:middle line:84% AUDIENCE: But it still sounds like an "s." 00:05:57.530 --> 00:05:58.040 align:middle line:90% NORVIN RICHARDS: Yeah. 00:05:58.040 --> 00:05:58.880 align:middle line:90% Right, exactly. 00:05:58.880 --> 00:06:01.250 align:middle line:84% When we pronounce it, we pronounce 00:06:01.250 --> 00:06:02.790 align:middle line:90% it like that, I think, "blitz." 00:06:02.790 --> 00:06:08.370 align:middle line:84% Yeah, we spell it like that because we got it from German. 00:06:08.370 --> 00:06:09.812 align:middle line:90% Yeah? 00:06:09.812 --> 00:06:11.270 align:middle line:90% OK. 00:06:11.270 --> 00:06:12.800 align:middle line:90% So yeah, no. 00:06:12.800 --> 00:06:14.550 align:middle line:84% There aren't any English words like that. 00:06:14.550 --> 00:06:18.980 align:middle line:90% We don't have words like that. 00:06:18.980 --> 00:06:20.720 align:middle line:84% There aren't any words like "cat-z." 00:06:20.720 --> 00:06:23.540 align:middle line:84% That's not a possible English word. 00:06:23.540 --> 00:06:27.320 align:middle line:84% And so if we're willing to say that the plural starts out 00:06:27.320 --> 00:06:32.870 align:middle line:84% as a "z," well, then we can handle this part 00:06:32.870 --> 00:06:34.790 align:middle line:84% of the-- not Raquel's part, not the part 00:06:34.790 --> 00:06:37.362 align:middle line:84% about "brushes" or "plushes," where 00:06:37.362 --> 00:06:38.570 align:middle line:90% there's a strident before it. 00:06:38.570 --> 00:06:40.340 align:middle line:84% But putting the stridents aside, we 00:06:40.340 --> 00:06:44.850 align:middle line:84% can handle this fact that the plural suffix is sometimes 00:06:44.850 --> 00:06:46.100 align:middle line:90% a "z' and sometimes an "s." 00:06:46.100 --> 00:06:48.798 align:middle line:84% We can make that part of English phonology generally. 00:06:48.798 --> 00:06:50.840 align:middle line:84% We don't have to make any statements specifically 00:06:50.840 --> 00:06:54.710 align:middle line:84% about this suffix, which is pleasing. 00:06:54.710 --> 00:06:57.740 align:middle line:84% So maybe that's an argument for making that move. 00:06:57.740 --> 00:07:00.300 align:middle line:90% 00:07:00.300 --> 00:07:03.920 align:middle line:84% So yeah, hypothesis two, the plural is underlying a "z," 00:07:03.920 --> 00:07:06.018 align:middle line:84% and it devoices after voiceless consonants. 00:07:06.018 --> 00:07:08.060 align:middle line:84% That can be part of a general English fact, which 00:07:08.060 --> 00:07:11.720 align:middle line:84% is English doesn't have words that end in a voiceless sound 00:07:11.720 --> 00:07:12.970 align:middle line:90% followed by a voiced sound. 00:07:12.970 --> 00:07:16.100 align:middle line:90% 00:07:16.100 --> 00:07:19.760 align:middle line:84% Similarly, after a strident, we get 00:07:19.760 --> 00:07:23.600 align:middle line:84% schwa "z," "brushes," "messes," "latches," "mazes," 00:07:23.600 --> 00:07:26.810 align:middle line:84% "plushes," that's Raquel's example. 00:07:26.810 --> 00:07:31.700 align:middle line:84% And English doesn't have words that end in two stridents. 00:07:31.700 --> 00:07:34.610 align:middle line:84% So the plural of "plush" is "plushez." 00:07:34.610 --> 00:07:39.500 align:middle line:90% It isn't "plushz." 00:07:39.500 --> 00:07:43.470 align:middle line:84% But we generally don't have sequences like that. 00:07:43.470 --> 00:07:45.290 align:middle line:84% And actually, this is an instance of we 00:07:45.290 --> 00:07:48.080 align:middle line:84% don't have "z" after a voiceless sound. 00:07:48.080 --> 00:07:51.350 align:middle line:90% But we also don't have-- 00:07:51.350 --> 00:07:52.940 align:middle line:84% yeah, what would we be looking for? 00:07:52.940 --> 00:07:59.180 align:middle line:90% 00:07:59.180 --> 00:08:02.960 align:middle line:84% I'm trying to think of something that ends in "dz." 00:08:02.960 --> 00:08:05.360 align:middle line:84% Is this a real word, or is it just something people use 00:08:05.360 --> 00:08:06.235 align:middle line:90% in crossword puzzles? 00:08:06.235 --> 00:08:08.780 align:middle line:90% AUDIENCE: I think it's a tool. 00:08:08.780 --> 00:08:11.450 align:middle line:84% NORVIN RICHARDS: Tool, like an ax, right? 00:08:11.450 --> 00:08:15.320 align:middle line:84% So if I get to pronounce this like this, 00:08:15.320 --> 00:08:18.680 align:middle line:84% the plural of it can't be "adz-z," right? 00:08:18.680 --> 00:08:21.620 align:middle line:90% It's "adzez," I guess. 00:08:21.620 --> 00:08:22.670 align:middle line:90% Yes. 00:08:22.670 --> 00:08:25.037 align:middle line:84% Actually, I don't know why I went for obscure tools. 00:08:25.037 --> 00:08:27.620 align:middle line:84% I could have done "mazes," which we have up here on the board. 00:08:27.620 --> 00:08:28.220 align:middle line:90% OK, yeah. 00:08:28.220 --> 00:08:28.700 align:middle line:90% AUDIENCE: "Rose"? 00:08:28.700 --> 00:08:28.955 align:middle line:90% NORVIN RICHARDS: Sorry? 00:08:28.955 --> 00:08:29.780 align:middle line:90% AUDIENCE: "Rose." 00:08:29.780 --> 00:08:30.780 align:middle line:90% NORVIN RICHARDS: "Rose." 00:08:30.780 --> 00:08:34.039 align:middle line:84% "Rose" would have done even better, yes. 00:08:34.039 --> 00:08:35.030 align:middle line:90% "Rose." 00:08:35.030 --> 00:08:37.400 align:middle line:90% Yeah, plural is "roses." 00:08:37.400 --> 00:08:42.110 align:middle line:84% So English doesn't have words that end in multiple stridents, 00:08:42.110 --> 00:08:44.480 align:middle line:84% and so yeah, we're putting a schwa there. 00:08:44.480 --> 00:08:47.960 align:middle line:84% Not something that we have to state maybe specifically 00:08:47.960 --> 00:08:50.180 align:middle line:84% about this morpheme, it's just a general fact 00:08:50.180 --> 00:08:51.300 align:middle line:90% about English phonology. 00:08:51.300 --> 00:08:54.310 align:middle line:90% 00:08:54.310 --> 00:09:04.050 align:middle line:84% So yeah, we can state rules that say things like an obstruent-- 00:09:04.050 --> 00:09:06.550 align:middle line:84% I'm sorry, I haven't introduced you to the word "obstruent." 00:09:06.550 --> 00:09:17.020 align:middle line:84% Obstruent is a word that covers stops and fricatives. 00:09:17.020 --> 00:09:21.220 align:middle line:84% Obstruents are sounds that create a build up 00:09:21.220 --> 00:09:24.140 align:middle line:90% of pressure in the oral cavity. 00:09:24.140 --> 00:09:26.830 align:middle line:84% So stops and fricatives are obstruents. 00:09:26.830 --> 00:09:31.680 align:middle line:84% Liquids, and glides, and vowels, and what else, liquids, 00:09:31.680 --> 00:09:35.280 align:middle line:84% and glides, and vowels, and nasals are not obstruents. 00:09:35.280 --> 00:09:39.120 align:middle line:84% So oral stops and fricatives, those are obstruents. 00:09:39.120 --> 00:09:44.980 align:middle line:84% Obstruents become voiceless after something voiceless. 00:09:44.980 --> 00:09:48.450 align:middle line:84% So underlying "cat-z" becomes "cats." 00:09:48.450 --> 00:09:52.182 align:middle line:84% And we have another rule here that inserts a schwa, 00:09:52.182 --> 00:09:53.640 align:middle line:84% and this is really just an exercise 00:09:53.640 --> 00:09:55.890 align:middle line:84% in showing you what phonological rules can look like. 00:09:55.890 --> 00:09:58.590 align:middle line:84% That second rule inserts a schwa between two stridents 00:09:58.590 --> 00:09:59.530 align:middle line:90% at the end of a word. 00:09:59.530 --> 00:10:02.130 align:middle line:84% So "inserts a schwa," the way you 00:10:02.130 --> 00:10:05.520 align:middle line:84% say that with these rules is "nothingness becomes a schwa." 00:10:05.520 --> 00:10:09.660 align:middle line:84% Yeah, I always liked that way of describing insertion. 00:10:09.660 --> 00:10:12.180 align:middle line:84% So you have this place where there isn't anything and it 00:10:12.180 --> 00:10:16.050 align:middle line:84% turns into something, namely a schwa, between stridents, 00:10:16.050 --> 00:10:17.940 align:middle line:90% so "brush-z" becomes "brushez." 00:10:17.940 --> 00:10:19.620 align:middle line:84% So we can state these rules, but we 00:10:19.620 --> 00:10:22.620 align:middle line:84% can think of these rules as things 00:10:22.620 --> 00:10:26.940 align:middle line:84% that English does in order to enforce its general conditions 00:10:26.940 --> 00:10:31.000 align:middle line:84% on what words can look like in these particular cases. 00:10:31.000 --> 00:10:33.813 align:middle line:90% Does that make sense so far? 00:10:33.813 --> 00:10:34.980 align:middle line:90% Questions about any of that? 00:10:34.980 --> 00:10:38.800 align:middle line:90% 00:10:38.800 --> 00:10:42.890 align:middle line:84% Oh, yeah, and then, if the only goal-- 00:10:42.890 --> 00:10:45.600 align:middle line:84% sorry, I should have remembered I was going to do this here. 00:10:45.600 --> 00:10:49.490 align:middle line:84% If the only goal of these sound changes 00:10:49.490 --> 00:10:55.910 align:middle line:84% is to create words that obey general conditions 00:10:55.910 --> 00:11:00.560 align:middle line:84% on English phonology, then we might ask ourselves, 00:11:00.560 --> 00:11:02.190 align:middle line:84% suppose you start off with that word. 00:11:02.190 --> 00:11:04.357 align:middle line:84% What is that word I've got down there at the bottom? 00:11:04.357 --> 00:11:08.898 align:middle line:90% 00:11:08.898 --> 00:11:10.440 align:middle line:84% Forget about the plural suffix on it. 00:11:10.440 --> 00:11:11.700 align:middle line:90% What's the singular? 00:11:11.700 --> 00:11:12.450 align:middle line:90% AUDIENCE: "Brush"? 00:11:12.450 --> 00:11:13.090 align:middle line:90% NORVIN RICHARDS: "Brush." 00:11:13.090 --> 00:11:13.590 align:middle line:90% Yeah. 00:11:13.590 --> 00:11:17.850 align:middle line:84% So suppose we start with "brush" and we're going to add the "z." 00:11:17.850 --> 00:11:20.580 align:middle line:84% Why not change the "z" to an "s"? 00:11:20.580 --> 00:11:23.400 align:middle line:84% After all, it's right next to a voiceless sound, giving you 00:11:23.400 --> 00:11:25.260 align:middle line:90% "brushs." 00:11:25.260 --> 00:11:29.610 align:middle line:84% And then they insert a schwa, giving you "brushes." 00:11:29.610 --> 00:11:30.735 align:middle line:90% That's not what you do. 00:11:30.735 --> 00:11:33.630 align:middle line:90% 00:11:33.630 --> 00:11:36.540 align:middle line:84% And we could wonder, why not, particularly 00:11:36.540 --> 00:11:39.720 align:middle line:84% if I've posited the two rules that I've got here, right? 00:11:39.720 --> 00:11:43.410 align:middle line:84% So I've got a rule that is going to devoice 00:11:43.410 --> 00:11:45.270 align:middle line:84% the "z" if it's at the end of a word 00:11:45.270 --> 00:11:47.490 align:middle line:90% and after something voiceless. 00:11:47.490 --> 00:11:50.820 align:middle line:84% And I've got a rule that inserts schwa between stridents. 00:11:50.820 --> 00:11:53.340 align:middle line:84% If I applied that first rule first and then the second rule, 00:11:53.340 --> 00:11:55.230 align:middle line:90% I would get this consequence. 00:11:55.230 --> 00:11:57.895 align:middle line:90% Does that make sense? 00:11:57.895 --> 00:11:59.770 align:middle line:84% And there are at least two ways of thinking-- 00:11:59.770 --> 00:12:01.290 align:middle line:84% sorry, you're making hand gestures 00:12:01.290 --> 00:12:03.220 align:middle line:84% that suggest you can see where I'm going with this-- 00:12:03.220 --> 00:12:04.650 align:middle line:84% There are at least two ways of thinking about this. 00:12:04.650 --> 00:12:06.670 align:middle line:84% There's the mechanical way, which is to say, well, 00:12:06.670 --> 00:12:08.462 align:middle line:84% we're learning something about these rules. 00:12:08.462 --> 00:12:09.754 align:middle line:90% They don't apply in that order. 00:12:09.754 --> 00:12:11.962 align:middle line:84% You don't do the first rule and then the second rule. 00:12:11.962 --> 00:12:13.980 align:middle line:84% You do the second rule and then the first rule. 00:12:13.980 --> 00:12:17.740 align:middle line:84% That's one way to think about this. 00:12:17.740 --> 00:12:20.610 align:middle line:84% There's another way to think about it, though. 00:12:20.610 --> 00:12:22.370 align:middle line:84% Look, what we've been saying right along 00:12:22.370 --> 00:12:26.690 align:middle line:84% is these rules are things that English does in order 00:12:26.690 --> 00:12:31.430 align:middle line:84% to make these plurals obey general English rules about how 00:12:31.430 --> 00:12:32.870 align:middle line:90% sound can be put together. 00:12:32.870 --> 00:12:38.400 align:middle line:84% I'm waving my arms a lot because I'm avoiding a technical term. 00:12:38.400 --> 00:12:41.446 align:middle line:84% And that's dumb, so let me just teach you a technical term. 00:12:41.446 --> 00:12:47.000 align:middle line:84% Phonotactics, this is the conditions 00:12:47.000 --> 00:12:51.740 align:middle line:84% on how sounds can combine, so what the rules are for which 00:12:51.740 --> 00:12:54.470 align:middle line:90% sounds can be where in a word. 00:12:54.470 --> 00:12:57.350 align:middle line:84% And I've now several times in the course of this lecture, 00:12:57.350 --> 00:12:59.060 align:middle line:84% I've almost said the word "phonotactics," 00:12:59.060 --> 00:13:00.170 align:middle line:84% and then stopped and said, wait, I 00:13:00.170 --> 00:13:01.460 align:middle line:84% haven't taught them that word yet. 00:13:01.460 --> 00:13:03.210 align:middle line:84% So I've been saying things like "the rules 00:13:03.210 --> 00:13:05.030 align:middle line:90% for how words can combine." 00:13:05.030 --> 00:13:08.780 align:middle line:84% So now you know that word, too, "phonotactics." 00:13:08.780 --> 00:13:10.370 align:middle line:84% Rules of English phonotactics include 00:13:10.370 --> 00:13:17.990 align:middle line:84% things like "English words can't end in a voiced sound that's 00:13:17.990 --> 00:13:20.000 align:middle line:84% immediately preceded by a voiceless sound," 00:13:20.000 --> 00:13:24.530 align:middle line:84% or "English words can't end in two stridents in a row." 00:13:24.530 --> 00:13:27.260 align:middle line:84% So we have these general conditions 00:13:27.260 --> 00:13:29.360 align:middle line:84% on English phonotactics, these conditions 00:13:29.360 --> 00:13:30.620 align:middle line:90% on how sounds can combine. 00:13:30.620 --> 00:13:33.170 align:middle line:84% And those two rules, those two things 00:13:33.170 --> 00:13:35.330 align:middle line:84% that happen to sequences of sounds 00:13:35.330 --> 00:13:38.690 align:middle line:84% are ways of enforcing those general conditions 00:13:38.690 --> 00:13:40.100 align:middle line:90% on phonotactics. 00:13:40.100 --> 00:13:43.200 align:middle line:90% Does that make sense? 00:13:43.200 --> 00:13:46.070 align:middle line:84% If we think about it that way, that is, if we think, 00:13:46.070 --> 00:13:49.397 align:middle line:84% there is, on the one hand, the rules-- 00:13:49.397 --> 00:13:51.480 align:middle line:84% the descriptions of the things that are happening. 00:13:51.480 --> 00:13:53.790 align:middle line:84% And then there is, on the other hand, 00:13:53.790 --> 00:13:56.750 align:middle line:84% the reasons why these things are happening. 00:13:56.750 --> 00:13:58.370 align:middle line:90% Does that make sense? 00:13:58.370 --> 00:14:00.830 align:middle line:84% So the reasons why these things are happening 00:14:00.830 --> 00:14:03.560 align:middle line:84% are things like, English doesn't like two stridents 00:14:03.560 --> 00:14:05.030 align:middle line:90% at the end of a word. 00:14:05.030 --> 00:14:07.250 align:middle line:84% That's why that second rule is happening. 00:14:07.250 --> 00:14:09.440 align:middle line:84% That's what it's for, it's to stop that 00:14:09.440 --> 00:14:12.233 align:middle line:84% from happening in the case of the English plural. 00:14:12.233 --> 00:14:14.150 align:middle line:84% So we were going to draw this distinction then 00:14:14.150 --> 00:14:15.920 align:middle line:84% between phonotactics, which say there 00:14:15.920 --> 00:14:18.210 align:middle line:84% are some combinations of sounds that are bad, 00:14:18.210 --> 00:14:20.480 align:middle line:84% and things like these rules which say, 00:14:20.480 --> 00:14:22.460 align:middle line:84% and if you have these combinations of sounds, 00:14:22.460 --> 00:14:25.890 align:middle line:84% here's what you ought to do to fix the problem. 00:14:25.890 --> 00:14:29.570 align:middle line:84% If you think about it, if I just tell you, 00:14:29.570 --> 00:14:33.020 align:middle line:84% having a word that ends in two stridents is bad in English. 00:14:33.020 --> 00:14:36.930 align:middle line:90% So if I start off with "brush"-- 00:14:36.930 --> 00:14:37.430 align:middle line:90% oh, dear. 00:14:37.430 --> 00:14:39.110 align:middle line:84% Now I have to write an "r" upside down. 00:14:39.110 --> 00:14:39.980 align:middle line:90% That's always hard. 00:14:39.980 --> 00:14:42.500 align:middle line:90% 00:14:42.500 --> 00:14:44.060 align:middle line:84% If I start off with "brush" and we're 00:14:44.060 --> 00:14:47.090 align:middle line:84% saying that the plural is underlyingly 00:14:47.090 --> 00:14:51.230 align:middle line:84% a "z," what we in fact do is to introduce a schwa. 00:14:51.230 --> 00:14:54.830 align:middle line:90% 00:14:54.830 --> 00:14:56.952 align:middle line:84% And we think the reason that we're doing that is, 00:14:56.952 --> 00:14:58.910 align:middle line:84% well, we're going to make reference to the fact 00:14:58.910 --> 00:15:00.980 align:middle line:84% that English generally doesn't allow words 00:15:00.980 --> 00:15:02.480 align:middle line:90% to end in multiple stridents. 00:15:02.480 --> 00:15:04.730 align:middle line:84% But if our only goal is to make sure this doesn't 00:15:04.730 --> 00:15:06.860 align:middle line:84% end in multiple stridents, well, there 00:15:06.860 --> 00:15:10.340 align:middle line:84% are lots of ways we could fix that problem. 00:15:10.340 --> 00:15:13.250 align:middle line:90% We could get rid of a strident. 00:15:13.250 --> 00:15:18.890 align:middle line:84% The plural of "brush" could be "bruhz," or "brush." 00:15:18.890 --> 00:15:22.160 align:middle line:84% We could introduce something other than a schwa. 00:15:22.160 --> 00:15:23.720 align:middle line:84% We could introduce a different vowel. 00:15:23.720 --> 00:15:25.520 align:middle line:90% It could be "brush-eez." 00:15:25.520 --> 00:15:27.110 align:middle line:90% We could introduce a consonant. 00:15:27.110 --> 00:15:28.730 align:middle line:90% It could be "brushts." 00:15:28.730 --> 00:15:30.620 align:middle line:84% Actually, that's not so easy to say either. 00:15:30.620 --> 00:15:32.030 align:middle line:90% "Brush-dz." 00:15:32.030 --> 00:15:32.600 align:middle line:90% Yeah. 00:15:32.600 --> 00:15:34.475 align:middle line:84% We could introduce some other consonants that 00:15:34.475 --> 00:15:35.978 align:middle line:90% would work better than these. 00:15:35.978 --> 00:15:38.270 align:middle line:84% Heck, we could introduce lots of consonants and vowels. 00:15:38.270 --> 00:15:44.840 align:middle line:84% The plural of "brush" could be "brush-glors." 00:15:44.840 --> 00:15:47.090 align:middle line:84% You can just introduce a whole chunk of stuff in here. 00:15:47.090 --> 00:15:49.640 align:middle line:84% There's all kinds of stuff you could do to fix that problem. 00:15:49.640 --> 00:15:53.300 align:middle line:90% This is what you, in fact, do. 00:15:53.300 --> 00:15:56.810 align:middle line:84% So when we say we're going to cover 00:15:56.810 --> 00:15:58.970 align:middle line:84% these facts about the different allomorphs 00:15:58.970 --> 00:16:02.330 align:middle line:84% of the English plural with general English phonology, 00:16:02.330 --> 00:16:05.160 align:middle line:84% is what I said, we got ourselves halfway there. 00:16:05.160 --> 00:16:07.190 align:middle line:84% We came up with these generalizations 00:16:07.190 --> 00:16:08.882 align:middle line:90% about English phonotactics. 00:16:08.882 --> 00:16:11.090 align:middle line:84% English doesn't like words that end in two stridents, 00:16:11.090 --> 00:16:11.760 align:middle line:90% let's say. 00:16:11.760 --> 00:16:14.480 align:middle line:84% And then we also have to have this solution to the problem, 00:16:14.480 --> 00:16:16.640 align:middle line:90% insert a schwa. 00:16:16.640 --> 00:16:19.850 align:middle line:84% The account of what's happening in English plurals 00:16:19.850 --> 00:16:21.590 align:middle line:90% needs both of those components. 00:16:21.590 --> 00:16:23.930 align:middle line:90% Does that make sense? 00:16:23.930 --> 00:16:27.620 align:middle line:84% Some of you are looking at me as though I'm making sense. 00:16:27.620 --> 00:16:29.730 align:middle line:84% It's always hard because you're all wearing masks. 00:16:29.730 --> 00:16:31.220 align:middle line:84% So I appreciate it, you've spent all this time 00:16:31.220 --> 00:16:33.080 align:middle line:84% working out ways of communicating emotions 00:16:33.080 --> 00:16:34.892 align:middle line:84% just with your eyebrows, which is great. 00:16:34.892 --> 00:16:36.350 align:middle line:84% All of you are waving your eyebrows 00:16:36.350 --> 00:16:38.070 align:middle line:84% in a semaphore-like fashion at me. 00:16:38.070 --> 00:16:40.520 align:middle line:90% I appreciate it a lot. 00:16:40.520 --> 00:16:42.570 align:middle line:84% So getting back to this question, 00:16:42.570 --> 00:16:49.610 align:middle line:84% why not first devoice the "s," and then introduce the schwa? 00:16:49.610 --> 00:16:53.910 align:middle line:84% The answer to that might be, well, that first-- 00:16:53.910 --> 00:16:56.490 align:middle line:84% so I said, there are two ways of talking about the answer 00:16:56.490 --> 00:16:56.990 align:middle line:90% to that. 00:16:56.990 --> 00:16:59.198 align:middle line:84% One is the mechanical way, which says, no, that's not 00:16:59.198 --> 00:17:00.920 align:middle line:90% the order these rules apply in. 00:17:00.920 --> 00:17:03.020 align:middle line:84% That's the way we had talked about this before. 00:17:03.020 --> 00:17:04.940 align:middle line:84% We went through these Lardil cases, 00:17:04.940 --> 00:17:08.420 align:middle line:84% and I said, sure looks as though we need to be willing to say, 00:17:08.420 --> 00:17:10.670 align:middle line:84% there are all these things that happen to these words, 00:17:10.670 --> 00:17:12.680 align:middle line:84% but we need them to happen in order so 00:17:12.680 --> 00:17:14.359 align:middle line:90% that we get the right answers. 00:17:14.359 --> 00:17:16.384 align:middle line:90% And that works. 00:17:16.384 --> 00:17:19.062 align:middle line:90% 00:17:19.062 --> 00:17:21.020 align:middle line:84% And in this particular case, we could say that. 00:17:21.020 --> 00:17:21.812 align:middle line:90% It would also work. 00:17:21.812 --> 00:17:23.750 align:middle line:84% We could say, the reason you don't 00:17:23.750 --> 00:17:26.250 align:middle line:84% do what I've got up there on the board, the reason you don't 00:17:26.250 --> 00:17:28.369 align:middle line:84% first devoice the "z" and then introduce the schwa 00:17:28.369 --> 00:17:31.190 align:middle line:84% is that that's not the order the operations take place in. 00:17:31.190 --> 00:17:33.770 align:middle line:84% But maybe this way of talking in terms of phonotactics 00:17:33.770 --> 00:17:36.980 align:middle line:90% is also helpful here. 00:17:36.980 --> 00:17:39.470 align:middle line:84% So we could say to ourselves, the reason 00:17:39.470 --> 00:17:42.920 align:middle line:84% that you don't first do that first thing is that it 00:17:42.920 --> 00:17:46.100 align:middle line:90% doesn't solve all the problems. 00:17:46.100 --> 00:17:49.310 align:middle line:84% Inserting a schwa solves all the problems at once. 00:17:49.310 --> 00:17:53.150 align:middle line:84% By inserting a schwa, you now no longer 00:17:53.150 --> 00:17:54.680 align:middle line:84% have a word ending in two stridents, 00:17:54.680 --> 00:17:58.820 align:middle line:84% and you also no longer have a word 00:17:58.820 --> 00:18:01.490 align:middle line:84% ending in a voiced sound preceded by a voiceless sound. 00:18:01.490 --> 00:18:03.860 align:middle line:84% Because the "z" at the end is now preceded by a schwa, 00:18:03.860 --> 00:18:05.880 align:middle line:90% which is voiced. 00:18:05.880 --> 00:18:08.940 align:middle line:84% So maybe there's some kind of principle of minimal repair, 00:18:08.940 --> 00:18:12.390 align:middle line:84% maybe, that says, look around at the various things you can do. 00:18:12.390 --> 00:18:14.420 align:middle line:90% Do the most effective one. 00:18:14.420 --> 00:18:17.010 align:middle line:84% That could be a way to talk about this. 00:18:17.010 --> 00:18:20.600 align:middle line:84% We're going to need to look at more cases to try to decide. 00:18:20.600 --> 00:18:24.020 align:middle line:84% But the point of going through the English plural 00:18:24.020 --> 00:18:26.900 align:middle line:84% in this level of detail was to introduce you to this idea. 00:18:26.900 --> 00:18:28.790 align:middle line:84% And having introduced you to this idea, 00:18:28.790 --> 00:18:31.610 align:middle line:84% we're now going to spend some time with this idea, this idea 00:18:31.610 --> 00:18:33.860 align:middle line:84% that when we're looking at a phonological problem, 00:18:33.860 --> 00:18:36.830 align:middle line:84% it's useful to think about it in these two halves. 00:18:36.830 --> 00:18:40.460 align:middle line:84% There's, on the one hand, what's the problem? 00:18:40.460 --> 00:18:42.560 align:middle line:90% Why are sounds changing? 00:18:42.560 --> 00:18:44.750 align:middle line:84% And on the other hand, what are the sounds doing? 00:18:44.750 --> 00:18:45.770 align:middle line:90% What's the fix? 00:18:45.770 --> 00:18:47.070 align:middle line:90% What's the repair? 00:18:47.070 --> 00:18:49.310 align:middle line:84% And we want to think about both of those things. 00:18:49.310 --> 00:18:50.640 align:middle line:90% Yeah. 00:18:50.640 --> 00:18:52.830 align:middle line:84% So in English, we've talked about a couple of kinds 00:18:52.830 --> 00:18:54.780 align:middle line:84% of problems that English plurals can create, 00:18:54.780 --> 00:18:57.460 align:middle line:84% and now we're also talking about some specific repairs. 00:18:57.460 --> 00:18:59.520 align:middle line:84% And we raise questions for ourselves. 00:18:59.520 --> 00:19:00.420 align:middle line:90% Why those repairs? 00:19:00.420 --> 00:19:02.725 align:middle line:90% Why not other repairs? 00:19:02.725 --> 00:19:03.975 align:middle line:90% We'll come back to that issue. 00:19:03.975 --> 00:19:08.230 align:middle line:90% 00:19:08.230 --> 00:19:11.710 align:middle line:84% One consequence of all this is that we're 00:19:11.710 --> 00:19:15.190 align:middle line:84% able to apply the rules of English plural formation 00:19:15.190 --> 00:19:18.970 align:middle line:84% to words that don't have the phonology of English words. 00:19:18.970 --> 00:19:21.430 align:middle line:84% Here are two words from other languages. 00:19:21.430 --> 00:19:24.300 align:middle line:84% Anybody want to pronounce them for us? 00:19:24.300 --> 00:19:24.800 align:middle line:90% Yeah. 00:19:24.800 --> 00:19:26.008 align:middle line:90% AUDIENCE: "Bach" and "rouge." 00:19:26.008 --> 00:19:28.280 align:middle line:84% NORVIN RICHARDS: Yeah, "Bach" and "rouge." 00:19:28.280 --> 00:19:28.850 align:middle line:90% Yeah. 00:19:28.850 --> 00:19:32.990 align:middle line:84% So where "Bach," if I'm being unbearably pretentious 00:19:32.990 --> 00:19:36.020 align:middle line:84% and I want to demonstrate for you the fact that I speak some 00:19:36.020 --> 00:19:40.490 align:middle line:84% German, I might pronounce the name of the great German family 00:19:40.490 --> 00:19:45.230 align:middle line:84% of German composers, the Bachs, was a big family, 00:19:45.230 --> 00:19:47.705 align:middle line:84% generations of Bachs writing all kinds of things. 00:19:47.705 --> 00:19:50.270 align:middle line:90% 00:19:50.270 --> 00:19:52.130 align:middle line:84% I might pronounce it with a velar fricative. 00:19:52.130 --> 00:19:53.797 align:middle line:84% The standard English way to pronounce it 00:19:53.797 --> 00:19:56.297 align:middle line:84% is with a velar stop because we don't have velar fricatives. 00:19:56.297 --> 00:19:57.890 align:middle line:84% But if I feel like being pretentious, 00:19:57.890 --> 00:20:01.370 align:middle line:84% I'll pronounce it with a velar fricative, "Bach." 00:20:01.370 --> 00:20:06.070 align:middle line:84% And then I've already done this, but what's the plural? 00:20:06.070 --> 00:20:07.400 align:middle line:90% "Bachs," right? 00:20:07.400 --> 00:20:09.670 align:middle line:90% It's with an "s." 00:20:09.670 --> 00:20:10.810 align:middle line:90% Why is it with an "s"? 00:20:10.810 --> 00:20:15.790 align:middle line:84% Well, because the repairs that we've been talking about 00:20:15.790 --> 00:20:18.590 align:middle line:84% are repairs for specific problems. 00:20:18.590 --> 00:20:24.400 align:middle line:84% So by hypothesis, the plural of this starts off as this. 00:20:24.400 --> 00:20:28.060 align:middle line:84% And now you have a word that ends in a voiced sound preceded 00:20:28.060 --> 00:20:29.300 align:middle line:90% by a voiceless sound. 00:20:29.300 --> 00:20:32.740 align:middle line:84% And the standard English repair for that is to devoice, 00:20:32.740 --> 00:20:34.540 align:middle line:90% so we change this "z" to an "s." 00:20:34.540 --> 00:20:35.665 align:middle line:90% I get "Bachs." 00:20:35.665 --> 00:20:38.290 align:middle line:90% 00:20:38.290 --> 00:20:42.910 align:middle line:84% Shows us that what we've got in our heads is not just-- 00:20:42.910 --> 00:20:45.688 align:middle line:84% so we've been going around this fact for a while now. 00:20:45.688 --> 00:20:47.230 align:middle line:84% What we have in our heads is not just 00:20:47.230 --> 00:20:53.170 align:middle line:84% a list of English nouns in their singular and plural forms. 00:20:53.170 --> 00:20:57.220 align:middle line:84% What we have are rules, algorithms for making plurals. 00:20:57.220 --> 00:20:58.720 align:middle line:84% And you can apply those algorithms 00:20:58.720 --> 00:21:00.897 align:middle line:84% to words that are not English words that 00:21:00.897 --> 00:21:02.980 align:middle line:84% have sounds in them that we don't have in English, 00:21:02.980 --> 00:21:07.120 align:middle line:84% like "Bach," or "rouge," which is the French word for "red," 00:21:07.120 --> 00:21:09.700 align:middle line:84% but I guess has been borrowed into English as a name 00:21:09.700 --> 00:21:11.230 align:middle line:90% for a component of makeup. 00:21:11.230 --> 00:21:14.530 align:middle line:84% If you're talking about a bunch of different kinds of rouge, 00:21:14.530 --> 00:21:17.110 align:middle line:84% which of these various kinds of rouge do you like best? 00:21:17.110 --> 00:21:20.462 align:middle line:90% Well, I like those. 00:21:20.462 --> 00:21:21.670 align:middle line:90% What's the plural of "rouge"? 00:21:21.670 --> 00:21:24.560 align:middle line:90% 00:21:24.560 --> 00:21:25.810 align:middle line:90% "Rouges" [with schwa and "z"]. 00:21:25.810 --> 00:21:28.170 align:middle line:90% "I like those rouges over there. 00:21:28.170 --> 00:21:32.140 align:middle line:84% I like these lipsticks better than those rouges." 00:21:32.140 --> 00:21:33.230 align:middle line:90% What did I do here? 00:21:33.230 --> 00:21:34.210 align:middle line:90% I inserted a schwa. 00:21:34.210 --> 00:21:34.900 align:middle line:90% Why? 00:21:34.900 --> 00:21:37.863 align:middle line:84% Well, because this is a strident there. 00:21:37.863 --> 00:21:40.030 align:middle line:84% It doesn't matter that it's not an English strident. 00:21:40.030 --> 00:21:42.670 align:middle line:84% We don't have that, except in words that we borrowed 00:21:42.670 --> 00:21:45.000 align:middle line:90% from French, like this one. 00:21:45.000 --> 00:21:45.818 align:middle line:90% Yeah? 00:21:45.818 --> 00:21:47.490 align:middle line:90% OK. 00:21:47.490 --> 00:21:48.000 align:middle line:90% All right. 00:21:48.000 --> 00:21:51.510 align:middle line:90% 00:21:51.510 --> 00:21:54.660 align:middle line:84% So this has all been an attempt to-- whoops, I spelled the "r" 00:21:54.660 --> 00:21:56.832 align:middle line:90% right side up again. 00:21:56.832 --> 00:21:58.290 align:middle line:84% This has all been an attempt to get 00:21:58.290 --> 00:22:03.443 align:middle line:84% you to take seriously the idea that it's useful when you're 00:22:03.443 --> 00:22:04.860 align:middle line:84% looking at a phonological problem, 00:22:04.860 --> 00:22:07.620 align:middle line:84% a phonological process, to ask yourself both, 00:22:07.620 --> 00:22:09.480 align:middle line:90% what is happening? 00:22:09.480 --> 00:22:10.800 align:middle line:90% So what has changed? 00:22:10.800 --> 00:22:14.400 align:middle line:84% What were the original things that I was combining? 00:22:14.400 --> 00:22:16.770 align:middle line:84% Let's say if I was adding an affix to a word, 00:22:16.770 --> 00:22:18.480 align:middle line:90% and what has happened to them? 00:22:18.480 --> 00:22:19.860 align:middle line:90% What form do they have now? 00:22:19.860 --> 00:22:20.940 align:middle line:90% Ask yourself that. 00:22:20.940 --> 00:22:23.070 align:middle line:90% And also ask, why? 00:22:23.070 --> 00:22:25.530 align:middle line:84% So what general properties of the language 00:22:25.530 --> 00:22:28.350 align:middle line:84% are we trying to enforce with this sound change? 00:22:28.350 --> 00:22:29.583 align:middle line:90% Yes? 00:22:29.583 --> 00:22:31.000 align:middle line:84% AUDIENCE: I was thinking about how 00:22:31.000 --> 00:22:33.250 align:middle line:84% you were saying that certain sound combinations just 00:22:33.250 --> 00:22:34.340 align:middle line:90% don't sound right. 00:22:34.340 --> 00:22:37.120 align:middle line:84% And I was thinking about how if you didn't know what part 00:22:37.120 --> 00:22:41.402 align:middle line:84% of speech the word "pore" is, then if you were like, OK, 00:22:41.402 --> 00:22:43.860 align:middle line:84% I'm going to pluralize it, then it would sound weird to say 00:22:43.860 --> 00:22:44.550 align:middle line:90% "por-s." 00:22:44.550 --> 00:22:46.050 align:middle line:84% But then if you knew it was supposed 00:22:46.050 --> 00:22:48.240 align:middle line:84% to be an adjective, then "porous" would actually 00:22:48.240 --> 00:22:50.520 align:middle line:90% sound kind of fine. 00:22:50.520 --> 00:22:55.820 align:middle line:84% So maybe there's some element of more structure 00:22:55.820 --> 00:22:58.840 align:middle line:84% that needs to be known to define what sounds good or weird? 00:22:58.840 --> 00:23:01.567 align:middle line:84% NORVIN RICHARDS: I think I might see what you mean. 00:23:01.567 --> 00:23:03.400 align:middle line:84% What was the word that you started off with? 00:23:03.400 --> 00:23:03.890 align:middle line:90% AUDIENCE: "Pore." 00:23:03.890 --> 00:23:04.890 align:middle line:90% NORVIN RICHARDS: "Pore." 00:23:04.890 --> 00:23:08.610 align:middle line:90% You mean that word? 00:23:08.610 --> 00:23:09.110 align:middle line:90% Yeah. 00:23:09.110 --> 00:23:12.410 align:middle line:84% And so we have those in our skin, little holes in our skin. 00:23:12.410 --> 00:23:15.290 align:middle line:90% And the plural is that, right? 00:23:15.290 --> 00:23:16.550 align:middle line:90% Yeah. 00:23:16.550 --> 00:23:17.310 align:middle line:90% Oh, I see. 00:23:17.310 --> 00:23:26.057 align:middle line:84% But you're saying-- so yeah, there's an adjective, "porous." 00:23:26.057 --> 00:23:27.390 align:middle line:90% Is that what you're thinking of? 00:23:27.390 --> 00:23:28.040 align:middle line:90% Yeah. 00:23:28.040 --> 00:23:37.850 align:middle line:84% And yes, so I'm trying to think about whether I can think 00:23:37.850 --> 00:23:40.115 align:middle line:84% of any adjectives that end with "-orz." 00:23:40.115 --> 00:23:44.550 align:middle line:90% 00:23:44.550 --> 00:23:47.197 align:middle line:84% I have "pores" on the brain, and "s'mores," and "floors," 00:23:47.197 --> 00:23:49.530 align:middle line:84% though everything I'm thinking of off the top of my head 00:23:49.530 --> 00:23:52.580 align:middle line:90% is plural. 00:23:52.580 --> 00:23:59.900 align:middle line:84% There are some generalizations about phonological differences 00:23:59.900 --> 00:24:06.290 align:middle line:84% sometimes between, occasionally, under certain circumstances, 00:24:06.290 --> 00:24:09.720 align:middle line:84% things of different syntactic categories. 00:24:09.720 --> 00:24:12.170 align:middle line:84% So there are languages that draw distinctions 00:24:12.170 --> 00:24:16.070 align:middle line:84% between, say, nouns and verbs, or nouns and adjectives. 00:24:16.070 --> 00:24:18.860 align:middle line:84% And nouns have a particular phonological signature, 00:24:18.860 --> 00:24:20.990 align:middle line:84% and adjectives have a different one. 00:24:20.990 --> 00:24:24.170 align:middle line:84% We'll actually get a chance to talk about an example like that 00:24:24.170 --> 00:24:28.530 align:middle line:84% when we get to Japanese, which has something like that. 00:24:28.530 --> 00:24:32.150 align:middle line:90% It's common, for example, for-- 00:24:32.150 --> 00:24:32.850 align:middle line:90% well, yeah. 00:24:32.850 --> 00:24:33.795 align:middle line:90% I'll leave it there. 00:24:33.795 --> 00:24:36.170 align:middle line:84% When we get to Japanese, we'll see something a little bit 00:24:36.170 --> 00:24:39.150 align:middle line:90% like that. 00:24:39.150 --> 00:24:42.290 align:middle line:84% And so maybe, getting back to your original comment, 00:24:42.290 --> 00:24:46.455 align:middle line:84% your original comment was, I'm talking as though what makes 00:24:46.455 --> 00:24:48.080 align:middle line:84% something look like a good English word 00:24:48.080 --> 00:24:50.600 align:middle line:84% is just going to be true of all English words. 00:24:50.600 --> 00:24:53.027 align:middle line:84% But maybe there are cases where we have to ask, 00:24:53.027 --> 00:24:54.860 align:middle line:84% does this look like a good English adjective 00:24:54.860 --> 00:24:56.330 align:middle line:90% or a good English verb? 00:24:56.330 --> 00:24:57.140 align:middle line:90% Yeah. 00:24:57.140 --> 00:25:00.230 align:middle line:84% And people have found things very vaguely like that, 00:25:00.230 --> 00:25:02.840 align:middle line:84% and I'll talk more about it when we get to Japanese. 00:25:02.840 --> 00:25:05.510 align:middle line:90% Yeah, nice point. 00:25:05.510 --> 00:25:08.013 align:middle line:84% Other points people wanted to make before we do Yawelmani? 00:25:08.013 --> 00:25:08.555 align:middle line:90% Yeah, Joseph? 00:25:08.555 --> 00:25:10.763 align:middle line:84% AUDIENCE: I'll ask you a question about transcription 00:25:10.763 --> 00:25:11.630 align:middle line:90% of "porous." 00:25:11.630 --> 00:25:16.370 align:middle line:84% If you use the power, would you use-- 00:25:16.370 --> 00:25:23.150 align:middle line:84% why would you use [INAUDIBLE] backwards C shape? 00:25:23.150 --> 00:25:26.535 align:middle line:84% NORVIN RICHARDS: Oh, you would use this? 00:25:26.535 --> 00:25:28.160 align:middle line:84% I think I would be tempted to pronounce 00:25:28.160 --> 00:25:30.640 align:middle line:90% that "pah," "ah," "pars." 00:25:30.640 --> 00:25:34.750 align:middle line:84% That's an "ah" sound, and so I guess 00:25:34.750 --> 00:25:36.910 align:middle line:84% that's why I wanted to spell it like this. 00:25:36.910 --> 00:25:40.180 align:middle line:84% I thought you were about to ask me why I'm not spelling it, 00:25:40.180 --> 00:25:44.860 align:middle line:84% say, like this, "porz" which I guess I could also do, 00:25:44.860 --> 00:25:49.507 align:middle line:90% like indicate the aspiration. 00:25:49.507 --> 00:25:52.090 align:middle line:84% I'm going to quickly erase this before anybody has any worries 00:25:52.090 --> 00:25:53.155 align:middle line:90% along any of these lines. 00:25:53.155 --> 00:25:57.160 align:middle line:90% 00:25:57.160 --> 00:26:00.700 align:middle line:84% Oh, and this actually brings me to something. 00:26:00.700 --> 00:26:04.240 align:middle line:84% I was raised, for some reason, to talk about decisions 00:26:04.240 --> 00:26:09.370 align:middle line:84% like this, decisions about whether to spell aspiration 00:26:09.370 --> 00:26:14.050 align:middle line:84% on English voiceless stops, to talk about that 00:26:14.050 --> 00:26:17.020 align:middle line:84% as a choice between what I was taught 00:26:17.020 --> 00:26:20.050 align:middle line:84% to call a "loose" and a "tight" transcription. 00:26:20.050 --> 00:26:22.450 align:middle line:84% But there is a more general term, which turns out 00:26:22.450 --> 00:26:23.740 align:middle line:90% to be used a lot more. 00:26:23.740 --> 00:26:25.690 align:middle line:90% I don't know who did this to me. 00:26:25.690 --> 00:26:28.480 align:middle line:84% The terms that people use more commonly 00:26:28.480 --> 00:26:30.400 align:middle line:90% are "broad" and "narrow." 00:26:30.400 --> 00:26:33.203 align:middle line:84% One of the TAs pointed this out to me after last time. 00:26:33.203 --> 00:26:34.870 align:middle line:84% He came up and said, why are you calling 00:26:34.870 --> 00:26:35.983 align:middle line:90% them "tight" and "loose"? 00:26:35.983 --> 00:26:37.150 align:middle line:90% I'm like, because I'm weird. 00:26:37.150 --> 00:26:38.057 align:middle line:90% I'm sorry. 00:26:38.057 --> 00:26:39.640 align:middle line:84% So "broad" and "narrow" is what you'll 00:26:39.640 --> 00:26:42.520 align:middle line:84% see more often, if you look in the literature. 00:26:42.520 --> 00:26:44.770 align:middle line:84% So I'm glad you asked your question because it allowed 00:26:44.770 --> 00:26:48.108 align:middle line:90% me to come clean about this. 00:26:48.108 --> 00:26:49.650 align:middle line:84% Other questions about why I am weird? 00:26:49.650 --> 00:26:50.190 align:middle line:90% Yes? 00:26:50.190 --> 00:26:52.440 align:middle line:84% AUDIENCE: Which one is broad, and which one is narrow? 00:26:52.440 --> 00:26:54.120 align:middle line:90% NORVIN RICHARDS: I'm sorry. 00:26:54.120 --> 00:26:58.770 align:middle line:84% Narrow is when you are trying to represent every fact. 00:26:58.770 --> 00:27:01.860 align:middle line:84% Broad is when you leave out facts that are predictable. 00:27:01.860 --> 00:27:04.800 align:middle line:84% So a broad transcription of "tense" 00:27:04.800 --> 00:27:07.320 align:middle line:84% would leave out, say, the aspiration on the "t." 00:27:07.320 --> 00:27:10.050 align:middle line:84% Because an English voiceless stop in that position 00:27:10.050 --> 00:27:13.350 align:middle line:84% is going to be aspirated, whereas a narrow transcription 00:27:13.350 --> 00:27:16.523 align:middle line:84% is trying to come as close as it can to a spectrogram. 00:27:16.523 --> 00:27:17.940 align:middle line:84% It's going to represent everything 00:27:17.940 --> 00:27:20.030 align:middle line:90% about the speech signal. 00:27:20.030 --> 00:27:23.630 align:middle line:84% I was taught to call that tight, but apparently I 00:27:23.630 --> 00:27:26.260 align:middle line:90% was taught by strange people. 00:27:26.260 --> 00:27:27.388 align:middle line:90% Yeah. 00:27:27.388 --> 00:27:29.180 align:middle line:84% I knew that I was taught by strange people. 00:27:29.180 --> 00:27:31.280 align:middle line:84% I just didn't realize how strange. 00:27:31.280 --> 00:27:34.440 align:middle line:90% Any other questions about that? 00:27:34.440 --> 00:27:35.450 align:middle line:90% All right. 00:27:35.450 --> 00:27:36.080 align:middle line:90% Yawelmani. 00:27:36.080 --> 00:27:37.400 align:middle line:90% OK, so I'm sorry. 00:27:37.400 --> 00:27:40.110 align:middle line:84% Let me do the intro to this slide again. 00:27:40.110 --> 00:27:46.730 align:middle line:84% So what I just did by taking you through the English plural 00:27:46.730 --> 00:27:49.143 align:middle line:84% forms was to try to introduce you to the idea, 00:27:49.143 --> 00:27:50.810 align:middle line:84% get you used to the idea which we're now 00:27:50.810 --> 00:27:53.270 align:middle line:84% about to see in action, again, in a different place. 00:27:53.270 --> 00:27:55.550 align:middle line:84% The idea that when you are looking 00:27:55.550 --> 00:27:58.130 align:middle line:84% at a phonological problem, it is useful to think 00:27:58.130 --> 00:28:01.520 align:middle line:84% both about what exactly is happening, 00:28:01.520 --> 00:28:03.060 align:middle line:84% what sound changes are happening, 00:28:03.060 --> 00:28:05.930 align:middle line:84% so in English, here we are inserting schwas or devoicing 00:28:05.930 --> 00:28:06.630 align:middle line:90% consonants. 00:28:06.630 --> 00:28:08.480 align:middle line:84% That's one kind of thing that's happening. 00:28:08.480 --> 00:28:09.870 align:middle line:90% It's useful to think about that. 00:28:09.870 --> 00:28:11.880 align:middle line:84% And it is also useful to ask yourself, 00:28:11.880 --> 00:28:13.280 align:middle line:90% why are these things happening? 00:28:13.280 --> 00:28:16.580 align:middle line:84% Can I think of this as a way of enforcing a general pattern 00:28:16.580 --> 00:28:18.680 align:middle line:90% in the language? 00:28:18.680 --> 00:28:22.230 align:middle line:84% It's useful to think about both of those things. 00:28:22.230 --> 00:28:25.160 align:middle line:84% So I'm about to show you another example of that in action. 00:28:25.160 --> 00:28:28.100 align:middle line:84% And let me just spoil the surprise for you right now. 00:28:28.100 --> 00:28:30.770 align:middle line:84% If there's anybody here who hates spoilers in movies, 00:28:30.770 --> 00:28:33.740 align:middle line:84% you should cover your ears or something. 00:28:33.740 --> 00:28:37.260 align:middle line:84% I'm going to show you a bunch of data from a language originally 00:28:37.260 --> 00:28:38.510 align:middle line:90% spoken in Northern California. 00:28:38.510 --> 00:28:39.843 align:middle line:90% I believe it's no longer spoken. 00:28:39.843 --> 00:28:41.990 align:middle line:90% It was called Yawelmani. 00:28:41.990 --> 00:28:46.760 align:middle line:84% And it was a language that had a general ban on having 00:28:46.760 --> 00:28:49.310 align:middle line:90% three consonants in a row. 00:28:49.310 --> 00:28:51.410 align:middle line:84% Now we're going to see that general ban 00:28:51.410 --> 00:28:54.200 align:middle line:84% in a bunch of different places with different fixes 00:28:54.200 --> 00:28:55.770 align:middle line:90% in different places. 00:28:55.770 --> 00:29:00.950 align:middle line:84% So the recognition that these different things that Yawelmani 00:29:00.950 --> 00:29:02.900 align:middle line:84% is doing under different circumstances 00:29:02.900 --> 00:29:07.100 align:middle line:84% are all instances of this ban was kind of a discovery, 00:29:07.100 --> 00:29:08.812 align:middle line:84% and people saying, yeah, it's doing this, 00:29:08.812 --> 00:29:10.520 align:middle line:84% and it's doing this, and it's doing this. 00:29:10.520 --> 00:29:12.380 align:middle line:84% And what all of those things achieve 00:29:12.380 --> 00:29:15.350 align:middle line:84% is avoiding sequences of three consonants in a row. 00:29:15.350 --> 00:29:18.215 align:middle line:84% So you guys have an advantage on the phonologists 00:29:18.215 --> 00:29:22.550 align:middle line:84% and the Amerindianists who were studying Yawelmani. 00:29:22.550 --> 00:29:24.770 align:middle line:84% They were mostly concentrating on what's 00:29:24.770 --> 00:29:26.300 align:middle line:84% happening in this form of the verb, what's happening 00:29:26.300 --> 00:29:27.360 align:middle line:90% in that form of the verb. 00:29:27.360 --> 00:29:27.860 align:middle line:90% Yeah. 00:29:27.860 --> 00:29:31.102 align:middle line:84% But you have the advantage that you have seen this slide. 00:29:31.102 --> 00:29:32.810 align:middle line:84% So remember what this slide says, and now 00:29:32.810 --> 00:29:34.100 align:middle line:90% let's look at some Yawelmani. 00:29:34.100 --> 00:29:36.620 align:middle line:90% 00:29:36.620 --> 00:29:41.420 align:middle line:84% Here are a bunch of Yawelmani verbs in their future forms. 00:29:41.420 --> 00:29:43.610 align:middle line:84% I don't speak Yawelmani so I won't 00:29:43.610 --> 00:29:44.910 align:middle line:90% try to pronounce these for you. 00:29:44.910 --> 00:29:47.040 align:middle line:84% But you can see that there's a suffix here, 00:29:47.040 --> 00:29:50.120 align:middle line:84% which is spelled E-N. And that suffix 00:29:50.120 --> 00:29:51.800 align:middle line:84% is the future tense suffix in Yawelmani. 00:29:51.800 --> 00:29:54.390 align:middle line:90% So far, so good? 00:29:54.390 --> 00:29:58.380 align:middle line:90% Now here's the gerund. 00:29:58.380 --> 00:29:59.857 align:middle line:84% The gerund is a form of the verb. 00:29:59.857 --> 00:30:02.190 align:middle line:84% Doesn't matter what the gerund is, but just so you know, 00:30:02.190 --> 00:30:04.020 align:middle line:84% a gerund is a form of the verb that 00:30:04.020 --> 00:30:07.360 align:middle line:84% allows it to be used as a noun, turns it into a noun. 00:30:07.360 --> 00:30:10.382 align:middle line:84% So when you're saying things in English like "swimming is fun," 00:30:10.382 --> 00:30:12.090 align:middle line:84% what you've done is take the verb "swim," 00:30:12.090 --> 00:30:14.460 align:middle line:84% and by adding "-ing," you've turned it into a gerund, 00:30:14.460 --> 00:30:16.835 align:middle line:84% that is a noun, the kind of thing that can be the subject 00:30:16.835 --> 00:30:19.680 align:middle line:84% of a sentence like "swimming is fun." 00:30:19.680 --> 00:30:22.000 align:middle line:84% These are gerunds in Yawelmani, and as you can see, 00:30:22.000 --> 00:30:23.910 align:middle line:84% Yawelmani gerunds involve a suffix, "-taw." 00:30:23.910 --> 00:30:28.350 align:middle line:90% 00:30:28.350 --> 00:30:33.120 align:middle line:84% Except in that last sequence of verbs, "sing," "pulverize," 00:30:33.120 --> 00:30:36.000 align:middle line:84% and "fight," you don't just add "-taw." 00:30:36.000 --> 00:30:39.090 align:middle line:84% You also insert a vowel, the vowel 00:30:39.090 --> 00:30:40.890 align:middle line:84% that I'll pronounce "ee", the vowel that's 00:30:40.890 --> 00:30:45.840 align:middle line:84% written with the letter "i," inside the original verb. 00:30:45.840 --> 00:30:49.230 align:middle line:84% So you don't get "ilk-taw," you get "ilik-taw." 00:30:49.230 --> 00:30:53.780 align:middle line:84% And you don't get "logw-taw," you get "logiw-taw." 00:30:53.780 --> 00:30:55.340 align:middle line:84% And because you saw the last slide, 00:30:55.340 --> 00:30:57.612 align:middle line:84% you know why Yawelmani is doing this. 00:30:57.612 --> 00:30:59.570 align:middle line:84% Yawelmani is doing this because it hates having 00:30:59.570 --> 00:31:01.250 align:middle line:90% three consonants in a row. 00:31:01.250 --> 00:31:06.167 align:middle line:84% So it's OK to add "-taw" to "mut" (swear) or "xat" (eat). 00:31:06.167 --> 00:31:08.000 align:middle line:84% I said I wouldn't try to pronounce Yawelmani 00:31:08.000 --> 00:31:09.120 align:middle line:90% but I guess I lied. 00:31:09.120 --> 00:31:11.120 align:middle line:84% Because there you're just putting two consonants 00:31:11.120 --> 00:31:12.660 align:middle line:90% in a row, and that's fine. 00:31:12.660 --> 00:31:14.330 align:middle line:84% But "sing," "pulverize," "fight," 00:31:14.330 --> 00:31:16.520 align:middle line:84% those verbs underlyingly end in two consonants. 00:31:16.520 --> 00:31:17.510 align:middle line:84% And if you add "-taw" after them, 00:31:17.510 --> 00:31:18.800 align:middle line:84% you're going to have three consonants in a row, 00:31:18.800 --> 00:31:19.730 align:middle line:90% and that's bad. 00:31:19.730 --> 00:31:21.772 align:middle line:84% Yawelmani doesn't like three consonants in a row, 00:31:21.772 --> 00:31:25.145 align:middle line:90% so you introduce this vowel. 00:31:25.145 --> 00:31:29.840 align:middle line:84% Now I just said that's why Yawelmani is doing this. 00:31:29.840 --> 00:31:32.720 align:middle line:84% At this stage, you could just look at this and go, well, no. 00:31:32.720 --> 00:31:33.860 align:middle line:90% Surely you're overreacting. 00:31:33.860 --> 00:31:36.995 align:middle line:90% Look, there's a rule. 00:31:36.995 --> 00:31:38.870 align:middle line:84% If you're forming the gerund, you add "-taw," 00:31:38.870 --> 00:31:41.120 align:middle line:84% and if the verb ends in two consonants, 00:31:41.120 --> 00:31:44.330 align:middle line:84% you insert a vowel between the two consonants as well. 00:31:44.330 --> 00:31:47.010 align:middle line:90% But wait, there's more. 00:31:47.010 --> 00:31:48.800 align:middle line:90% So here's one rule. 00:31:48.800 --> 00:31:51.560 align:middle line:84% Nothingness changes into an "i," so you 00:31:51.560 --> 00:31:56.150 align:middle line:84% insert an "i" between the first and second consonants 00:31:56.150 --> 00:31:58.160 align:middle line:84% when you have three consonants in a row. 00:31:58.160 --> 00:32:00.610 align:middle line:90% That's what that rule says. 00:32:00.610 --> 00:32:02.140 align:middle line:90% Here's another rule. 00:32:02.140 --> 00:32:03.860 align:middle line:90% Here's the desiderative. 00:32:03.860 --> 00:32:06.640 align:middle line:84% So you can take a verb in Yawelmani and add a suffix, 00:32:06.640 --> 00:32:10.690 align:middle line:84% "-hatin," that gives you a new verb that means, 00:32:10.690 --> 00:32:16.540 align:middle line:84% "want to (verb)," so "want to know," or "want to sink," 00:32:16.540 --> 00:32:19.420 align:middle line:84% refraining from wondering why you would ever want to sink. 00:32:19.420 --> 00:32:23.190 align:middle line:90% 00:32:23.190 --> 00:32:25.680 align:middle line:84% The desiderative suffix is usually "-hatin," 00:32:25.680 --> 00:32:28.530 align:middle line:84% except if you're adding it to a verb that ends in two 00:32:28.530 --> 00:32:30.960 align:middle line:84% consonants, like "speak" or "lift," well, 00:32:30.960 --> 00:32:34.620 align:middle line:84% then the desiderative suffix is no longer "-hatin." 00:32:34.620 --> 00:32:35.730 align:middle line:90% It's just "-atin." 00:32:35.730 --> 00:32:39.000 align:middle line:90% You take off the "h." 00:32:39.000 --> 00:32:39.690 align:middle line:90% OK. 00:32:39.690 --> 00:32:42.240 align:middle line:84% So here's another thing that Yawelmani does. 00:32:42.240 --> 00:32:45.810 align:middle line:84% If you've got two consonants and then an "h," 00:32:45.810 --> 00:32:48.510 align:middle line:90% you get rid of the "h." 00:32:48.510 --> 00:32:50.460 align:middle line:84% Again, we have two rules, both fine rules. 00:32:50.460 --> 00:32:52.760 align:middle line:90% They both work. 00:32:52.760 --> 00:32:55.160 align:middle line:84% But by just stating them, by saying, here are two rules, 00:32:55.160 --> 00:32:56.827 align:middle line:84% here are two things that Yawelmani does, 00:32:56.827 --> 00:32:59.450 align:middle line:84% we're obscuring something we feel like. 00:32:59.450 --> 00:33:01.443 align:middle line:84% Yawelmani does both of these things, 00:33:01.443 --> 00:33:03.110 align:middle line:84% but they both have the same consequence. 00:33:03.110 --> 00:33:05.112 align:middle line:84% They are two ways to avoid strings 00:33:05.112 --> 00:33:06.320 align:middle line:90% of three consonants in a row. 00:33:06.320 --> 00:33:08.450 align:middle line:90% Yes? 00:33:08.450 --> 00:33:10.672 align:middle line:84% AUDIENCE: Is this three consonants [INAUDIBLE]?? 00:33:10.672 --> 00:33:13.980 align:middle line:90% 00:33:13.980 --> 00:33:19.050 align:middle line:90% I see [INAUDIBLE] double Ls. 00:33:19.050 --> 00:33:21.530 align:middle line:90% I don't really see that. 00:33:21.530 --> 00:33:23.435 align:middle line:84% Is that actually just two consonants 00:33:23.435 --> 00:33:24.620 align:middle line:90% sounds that are [INAUDIBLE] 00:33:24.620 --> 00:33:25.640 align:middle line:90% NORVIN RICHARDS: Yeah. 00:33:25.640 --> 00:33:27.710 align:middle line:90% So it's two, yes. 00:33:27.710 --> 00:33:30.050 align:middle line:90% It is a geminate "l." 00:33:30.050 --> 00:33:33.920 align:middle line:84% I say this out of my vast, vast knowledge of Yawelmani. 00:33:33.920 --> 00:33:35.457 align:middle line:90% I'm being ironic. 00:33:35.457 --> 00:33:37.040 align:middle line:84% I don't know anything about Yawelmani, 00:33:37.040 --> 00:33:38.930 align:middle line:90% but this is a geminate "l." 00:33:38.930 --> 00:33:41.270 align:middle line:84% It's an "l" sound that's held for two beats. 00:33:41.270 --> 00:33:44.120 align:middle line:84% So yes, we need that to count as two consonants. 00:33:44.120 --> 00:33:45.690 align:middle line:84% You're raising a really good point. 00:33:45.690 --> 00:33:46.820 align:middle line:84% We have to ask ourselves, when we 00:33:46.820 --> 00:33:48.987 align:middle line:84% say it's bad to have three consonants in a row, what 00:33:48.987 --> 00:33:50.550 align:middle line:90% counts as a consonant? 00:33:50.550 --> 00:33:52.460 align:middle line:90% This had better be one. 00:33:52.460 --> 00:33:54.400 align:middle line:90% Yeah. 00:33:54.400 --> 00:33:56.460 align:middle line:90% Good question. 00:33:56.460 --> 00:33:58.170 align:middle line:90% All right, more Yawelmani. 00:33:58.170 --> 00:34:02.670 align:middle line:90% 00:34:02.670 --> 00:34:06.840 align:middle line:84% Here is a passive suffix, spelled H-N-E-L. 00:34:06.840 --> 00:34:08.500 align:middle line:90% I won't attempt to pronounce it. 00:34:08.500 --> 00:34:10.830 align:middle line:84% So here we've got the verb "to be tied" 00:34:10.830 --> 00:34:12.330 align:middle line:90% and the verb "to be hit." 00:34:12.330 --> 00:34:13.980 align:middle line:84% So you can take the verb "to tie," 00:34:13.980 --> 00:34:17.050 align:middle line:84% "t'ik'e," which has two of those cool ejectives, 00:34:17.050 --> 00:34:23.050 align:middle line:84% and you can add this passive suffix to it, H-N-E-L. 00:34:23.050 --> 00:34:26.659 align:middle line:84% Maybe all of you can imagine what's going to happen next. 00:34:26.659 --> 00:34:29.710 align:middle line:84% If we add this suffix to something that already ends 00:34:29.710 --> 00:34:32.980 align:middle line:84% in a consonant, well, then we get rid of the H. 00:34:32.980 --> 00:34:36.438 align:middle line:84% So "to be helped" or "to be held under the arm," 00:34:36.438 --> 00:34:38.230 align:middle line:84% those are two verbs that end in consonants, 00:34:38.230 --> 00:34:41.889 align:middle line:84% and now the suffix is just "-nel." 00:34:41.889 --> 00:34:44.440 align:middle line:84% So in the last slide, we saw that "h" goes away 00:34:44.440 --> 00:34:46.389 align:middle line:84% if it's preceded by two consonants. 00:34:46.389 --> 00:34:48.909 align:middle line:84% In this slide, we see that "h" goes away 00:34:48.909 --> 00:34:51.760 align:middle line:84% if it has consonants on either side of it. 00:34:51.760 --> 00:34:53.810 align:middle line:90% Fine, we can say it that way. 00:34:53.810 --> 00:34:57.162 align:middle line:84% We can state different rules for all of these morphemes. 00:34:57.162 --> 00:34:59.620 align:middle line:84% And in fact, the first people who worked on Yawelmani money 00:34:59.620 --> 00:35:00.237 align:middle line:90% did just that. 00:35:00.237 --> 00:35:02.320 align:middle line:84% They were like, OK, the passive suffix is "-hnel," 00:35:02.320 --> 00:35:04.780 align:middle line:84% unless the verb ends in a consonant, 00:35:04.780 --> 00:35:05.860 align:middle line:90% and then it's "-nel." 00:35:05.860 --> 00:35:07.960 align:middle line:84% The desiderative suffix is "-hatin," 00:35:07.960 --> 00:35:10.490 align:middle line:84% unless the verb ends in two consonants, 00:35:10.490 --> 00:35:12.430 align:middle line:90% and then it's "-atin." 00:35:12.430 --> 00:35:17.360 align:middle line:84% You can do that but you're missing a generalization, which 00:35:17.360 --> 00:35:19.460 align:middle line:84% is Yawelmani is a language that hates having 00:35:19.460 --> 00:35:22.860 align:middle line:90% three consonants in a row. 00:35:22.860 --> 00:35:25.017 align:middle line:84% So this is like the distinction that we 00:35:25.017 --> 00:35:27.350 align:middle line:84% were drawing when we were talking about English plurals. 00:35:27.350 --> 00:35:29.142 align:middle line:84% When we were talking about English plurals, 00:35:29.142 --> 00:35:31.490 align:middle line:84% we said, yeah, English doesn't like 00:35:31.490 --> 00:35:34.790 align:middle line:84% words that end in two stridents, or words 00:35:34.790 --> 00:35:37.550 align:middle line:84% that end in a voiced sound preceded 00:35:37.550 --> 00:35:40.340 align:middle line:84% by a voiceless sounding, which doesn't allow words like that. 00:35:40.340 --> 00:35:43.910 align:middle line:84% And English has repairs, things it does, it devoices, 00:35:43.910 --> 00:35:45.587 align:middle line:90% or it introduces a schwa. 00:35:45.587 --> 00:35:47.420 align:middle line:84% We're seeing something similar in Yawelmani. 00:35:47.420 --> 00:35:47.920 align:middle line:90% Yeah? 00:35:47.920 --> 00:35:51.760 align:middle line:84% AUDIENCE: So in two examples, we had "-hatin" and "-nel," 00:35:51.760 --> 00:35:53.283 align:middle line:90% which were both [INAUDIBLE]. 00:35:53.283 --> 00:35:54.200 align:middle line:90% NORVIN RICHARDS: Yeah. 00:35:54.200 --> 00:36:03.670 align:middle line:84% AUDIENCE: [INAUDIBLE],, should we try to go through the sounds 00:36:03.670 --> 00:36:04.210 align:middle line:90% itself? 00:36:04.210 --> 00:36:07.600 align:middle line:84% Or would this make someone who's trying 00:36:07.600 --> 00:36:11.050 align:middle line:84% to come up with the rules mad to consider something like, 00:36:11.050 --> 00:36:13.780 align:middle line:84% if "h" is part of the suffix, then 00:36:13.780 --> 00:36:16.300 align:middle line:84% drop the "h," instead of the position of "h," 00:36:16.300 --> 00:36:18.420 align:middle line:90% the three consonants? 00:36:18.420 --> 00:36:20.920 align:middle line:84% NORVIN RICHARDS: I'm sorry, can you ask your question again? 00:36:20.920 --> 00:36:22.087 align:middle line:90% I'm not sure I'm getting it. 00:36:22.087 --> 00:36:23.120 align:middle line:90% Can you just say again? 00:36:23.120 --> 00:36:25.840 align:middle line:84% AUDIENCE: So in "-hatin" and "-nel" example, 00:36:25.840 --> 00:36:30.810 align:middle line:84% the conclusion was about [INAUDIBLE] and the three 00:36:30.810 --> 00:36:31.605 align:middle line:90% consonants. 00:36:31.605 --> 00:36:34.280 align:middle line:90% 00:36:34.280 --> 00:36:37.130 align:middle line:84% Could it also be about whether "h" 00:36:37.130 --> 00:36:39.095 align:middle line:90% is part of the added suffix? 00:36:39.095 --> 00:36:40.220 align:middle line:90% NORVIN RICHARDS: Oh, I see. 00:36:40.220 --> 00:36:43.437 align:middle line:90% I see, I see. 00:36:43.437 --> 00:36:44.270 align:middle line:90% That's a nice point. 00:36:44.270 --> 00:36:51.070 align:middle line:84% So could you say, suffixes that begin with an "h" 00:36:51.070 --> 00:36:58.360 align:middle line:84% won't begin with an "h" if the "h" is going to create 00:36:58.360 --> 00:37:00.040 align:middle line:90% a sequence of three consonants? 00:37:00.040 --> 00:37:02.410 align:middle line:90% Yes. 00:37:02.410 --> 00:37:07.600 align:middle line:84% Yes, indeed, that covers everything 00:37:07.600 --> 00:37:11.110 align:middle line:84% I did with my last two rules, though it's not 00:37:11.110 --> 00:37:15.070 align:middle line:84% something that's easy to state with the rule formalism 00:37:15.070 --> 00:37:17.500 align:middle line:90% that I showed you. 00:37:17.500 --> 00:37:20.110 align:middle line:84% Maybe the response to that is, so we need a better rule 00:37:20.110 --> 00:37:22.090 align:middle line:90% formalism. 00:37:22.090 --> 00:37:26.440 align:middle line:90% But no, that's absolutely right. 00:37:26.440 --> 00:37:28.180 align:middle line:84% Really what you're asking, in a way, 00:37:28.180 --> 00:37:31.180 align:middle line:84% is, is it important that those examples 00:37:31.180 --> 00:37:33.820 align:middle line:84% involved in "h" that was at the beginning of a suffix? 00:37:33.820 --> 00:37:35.410 align:middle line:84% Could I give you any examples where 00:37:35.410 --> 00:37:40.030 align:middle line:84% there was a verb that ended in two consonants, let's say, 00:37:40.030 --> 00:37:42.850 align:middle line:90% of which the last was an "h"? 00:37:42.850 --> 00:37:45.850 align:middle line:84% And then when you add a suffix, the "h' will go away. 00:37:45.850 --> 00:37:48.250 align:middle line:84% That's what these rules are predicting. 00:37:48.250 --> 00:37:50.630 align:middle line:84% And your rule is predicting something else. 00:37:50.630 --> 00:37:52.530 align:middle line:84% And I don't know enough about Yawelmani 00:37:52.530 --> 00:37:54.493 align:middle line:84% to know whether there are examples that allow 00:37:54.493 --> 00:37:55.660 align:middle line:90% you to draw the distinction. 00:37:55.660 --> 00:37:58.035 align:middle line:84% But you're absolutely right, that's what you'd want to go 00:37:58.035 --> 00:38:00.280 align:middle line:84% do is to find out whether the suffix matters. 00:38:00.280 --> 00:38:03.380 align:middle line:84% AUDIENCE: So if you have a suffix 00:38:03.380 --> 00:38:06.263 align:middle line:84% which has "h" in the second place of a consonant-- 00:38:06.263 --> 00:38:08.680 align:middle line:84% NORVIN RICHARDS: That would be another good place to look, 00:38:08.680 --> 00:38:09.180 align:middle line:90% wouldn't it? 00:38:09.180 --> 00:38:13.010 align:middle line:84% AUDIENCE: --in that case, would the "h" drop, or [INAUDIBLE]?? 00:38:13.010 --> 00:38:16.520 align:middle line:84% NORVIN RICHARDS: So yeah, let's make up a Yawelmani consonant. 00:38:16.520 --> 00:38:20.870 align:middle line:84% If there were a consonant in H-E-L, 00:38:20.870 --> 00:38:25.190 align:middle line:84% I guess these rules predict that if you added that 00:38:25.190 --> 00:38:27.500 align:middle line:84% to a verb that ended in a consonant 00:38:27.500 --> 00:38:30.740 align:middle line:90% that the "h" would drop. 00:38:30.740 --> 00:38:33.350 align:middle line:84% And so that's what these rules predict. 00:38:33.350 --> 00:38:35.450 align:middle line:84% The imaginable rule that you're talking 00:38:35.450 --> 00:38:38.120 align:middle line:84% about that says "h" at the beginning of a suffix 00:38:38.120 --> 00:38:40.713 align:middle line:84% is particularly vulnerable, and you'll 00:38:40.713 --> 00:38:42.380 align:middle line:84% drop an "h" at the beginning of a suffix 00:38:42.380 --> 00:38:44.330 align:middle line:84% as a way of fixing this general rule. 00:38:44.330 --> 00:38:47.430 align:middle line:84% And then if you can't do that, then you insert a vowel. 00:38:47.430 --> 00:38:48.930 align:middle line:84% Maybe that's a way to talk about it. 00:38:48.930 --> 00:38:50.305 align:middle line:84% That makes a different prediction 00:38:50.305 --> 00:38:51.665 align:middle line:90% about this kind of suffix. 00:38:51.665 --> 00:38:54.290 align:middle line:84% I'm pretty sure there aren't any suffixes that have that shape. 00:38:54.290 --> 00:38:55.550 align:middle line:84% I don't know much about Yawelmani 00:38:55.550 --> 00:38:57.800 align:middle line:84% but I'm pretty sure there aren't any things like that. 00:38:57.800 --> 00:39:01.730 align:middle line:84% But all of us should now go find things on Yawelmani 00:39:01.730 --> 00:39:02.360 align:middle line:90% and read them. 00:39:02.360 --> 00:39:04.297 align:middle line:84% Yeah, so this stuff is all out there. 00:39:04.297 --> 00:39:05.630 align:middle line:90% It'd be interesting to find out. 00:39:05.630 --> 00:39:07.088 align:middle line:84% But you're right, this is exactly-- 00:39:07.088 --> 00:39:09.380 align:middle line:84% what you're doing is exactly what we want to do now. 00:39:09.380 --> 00:39:11.450 align:middle line:84% So I'm showing you one set of rules 00:39:11.450 --> 00:39:12.710 align:middle line:90% that would cover these data. 00:39:12.710 --> 00:39:14.690 align:middle line:84% And it's worth asking yourself, what other kinds of rules 00:39:14.690 --> 00:39:15.648 align:middle line:90% would cover these data? 00:39:15.648 --> 00:39:18.020 align:middle line:84% And how would we find out which of these rules 00:39:18.020 --> 00:39:19.040 align:middle line:90% are the right ones? 00:39:19.040 --> 00:39:21.260 align:middle line:84% And the only answer is the boring one, 00:39:21.260 --> 00:39:23.090 align:middle line:84% which is, go learn more about Yawelmani. 00:39:23.090 --> 00:39:25.790 align:middle line:84% But you're right, that's the first thing to do. 00:39:25.790 --> 00:39:27.830 align:middle line:84% AUDIENCE: Yeah, my main question now 00:39:27.830 --> 00:39:30.410 align:middle line:84% is more of the lines of what you said. 00:39:30.410 --> 00:39:32.940 align:middle line:84% This rule of formalism doesn't allow these kinds of rules, 00:39:32.940 --> 00:39:35.090 align:middle line:84% right, so do we just not want to [INAUDIBLE]?? 00:39:35.090 --> 00:39:35.510 align:middle line:90% NORVIN RICHARDS: No. 00:39:35.510 --> 00:39:36.560 align:middle line:90% Well, so we want to find out. 00:39:36.560 --> 00:39:37.925 align:middle line:84% Because if there are rules-- you're 00:39:37.925 --> 00:39:39.150 align:middle line:90% raising a really good point. 00:39:39.150 --> 00:39:41.210 align:middle line:84% If there are rules like that, then we 00:39:41.210 --> 00:39:45.080 align:middle line:84% want a rule formalism that allows us to state them. 00:39:45.080 --> 00:39:47.690 align:middle line:84% It's not like we would have to kill ourselves making that rule 00:39:47.690 --> 00:39:48.500 align:middle line:90% formalism either. 00:39:48.500 --> 00:39:50.150 align:middle line:84% So the rule formalism that I showed 00:39:50.150 --> 00:39:54.350 align:middle line:84% you has a symbol in it for a word boundary that 00:39:54.350 --> 00:39:56.630 align:middle line:84% says there are special things that 00:39:56.630 --> 00:39:59.450 align:middle line:90% happen at the end of a word. 00:39:59.450 --> 00:40:03.873 align:middle line:84% So there's this symbol that says, in English, 00:40:03.873 --> 00:40:06.290 align:middle line:84% if you have two stridents at the end of a word then you've 00:40:06.290 --> 00:40:09.410 align:middle line:84% done something wrong and you must introduce a schwa. 00:40:09.410 --> 00:40:13.850 align:middle line:84% No reason we couldn't introduce a symbol for morpheme boundary. 00:40:13.850 --> 00:40:17.870 align:middle line:84% And then we would use it to capture your generalization. 00:40:17.870 --> 00:40:20.618 align:middle line:84% In order to find out whether we need that for Yawelmani, 00:40:20.618 --> 00:40:22.160 align:middle line:84% well, we need to know more Yawelmani. 00:40:22.160 --> 00:40:24.743 align:middle line:84% But the question you're raising is exactly the right question, 00:40:24.743 --> 00:40:28.250 align:middle line:84% and it's how we find out whether we need devices like that. 00:40:28.250 --> 00:40:29.210 align:middle line:90% Yeah, cool. 00:40:29.210 --> 00:40:30.870 align:middle line:84% That's linguistics you're doing there. 00:40:30.870 --> 00:40:31.370 align:middle line:90% Yeah? 00:40:31.370 --> 00:40:33.162 align:middle line:84% AUDIENCE: What exactly are the apostrophes? 00:40:33.162 --> 00:40:35.090 align:middle line:84% NORVIN RICHARDS: Oh, good question. 00:40:35.090 --> 00:40:39.720 align:middle line:84% These particular apostrophes are markers of ejective stops. 00:40:39.720 --> 00:40:47.240 align:middle line:84% So the verb to hit is "tok-o" with an ejective "k." 00:40:47.240 --> 00:40:50.450 align:middle line:90% That's what those are. 00:40:50.450 --> 00:40:52.040 align:middle line:90% People remember ejective stops. 00:40:52.040 --> 00:40:54.350 align:middle line:84% All of you were ejectively stopping at me 00:40:54.350 --> 00:40:56.780 align:middle line:84% during the phonetics part of the class. 00:40:56.780 --> 00:40:58.970 align:middle line:84% This is where you make a glottal stop 00:40:58.970 --> 00:41:02.540 align:middle line:84% and express some air from your mouth during a stop 00:41:02.540 --> 00:41:07.220 align:middle line:84% by moving your larynx, shoving the air out that way, "ka." 00:41:07.220 --> 00:41:09.140 align:middle line:90% OK, cool. 00:41:09.140 --> 00:41:13.370 align:middle line:84% So to summarize then, we have these rules 00:41:13.370 --> 00:41:17.930 align:middle line:84% and this interesting discussion about whether they're 00:41:17.930 --> 00:41:18.650 align:middle line:90% the right rules. 00:41:18.650 --> 00:41:22.370 align:middle line:84% There are various ways of categorizing the rules. 00:41:22.370 --> 00:41:27.260 align:middle line:84% And what I'm suggesting here is that we 00:41:27.260 --> 00:41:29.210 align:middle line:84% might want to be willing to abstract away 00:41:29.210 --> 00:41:31.580 align:middle line:84% from the particular rules that I posited and say, 00:41:31.580 --> 00:41:34.910 align:middle line:84% Yawelmani doesn't like strings of three consonants in a row. 00:41:34.910 --> 00:41:37.490 align:middle line:84% And it has these general principles that say, 00:41:37.490 --> 00:41:40.610 align:middle line:84% you can introduce a vowel to break up sequences, 00:41:40.610 --> 00:41:41.917 align:middle line:90% and you can get rid of an "h." 00:41:41.917 --> 00:41:44.000 align:middle line:84% And the question that was getting raised over here 00:41:44.000 --> 00:41:46.880 align:middle line:84% was, is it just, you can always get rid of an "h"? 00:41:46.880 --> 00:41:48.470 align:middle line:84% Or is it, you can get rid of an "h," 00:41:48.470 --> 00:41:50.960 align:middle line:84% but only if it's the beginning of a suffix? 00:41:50.960 --> 00:41:53.510 align:middle line:84% Because that was the case in all of the examples 00:41:53.510 --> 00:41:56.180 align:middle line:84% that I showed you, to which, unfortunately, 00:41:56.180 --> 00:41:58.430 align:middle line:84% the only answer is, go ask someone who knows more 00:41:58.430 --> 00:42:00.440 align:middle line:84% Yawelmani than me, which is most people. 00:42:00.440 --> 00:42:01.470 align:middle line:90% No, that's not true. 00:42:01.470 --> 00:42:05.330 align:middle line:84% There are several people who know more Yawelmani than me. 00:42:05.330 --> 00:42:06.800 align:middle line:90% So do people see this? 00:42:06.800 --> 00:42:08.480 align:middle line:84% We're making the same move here that we 00:42:08.480 --> 00:42:10.040 align:middle line:84% made when we were talking about the English plurals. 00:42:10.040 --> 00:42:12.740 align:middle line:84% So we said, yeah, English has these particular things 00:42:12.740 --> 00:42:14.270 align:middle line:90% that it does, these repairs. 00:42:14.270 --> 00:42:17.210 align:middle line:84% And these repairs are there because of these conditions 00:42:17.210 --> 00:42:19.250 align:middle line:84% on English phonotactics, rules about how 00:42:19.250 --> 00:42:20.585 align:middle line:90% sounds can combine in English. 00:42:20.585 --> 00:42:21.710 align:middle line:90% Similar kind of thing here. 00:42:21.710 --> 00:42:24.260 align:middle line:84% There's a principle of Yawelmani phonotactics. 00:42:24.260 --> 00:42:25.880 align:middle line:84% Don't have three consonants in a row, 00:42:25.880 --> 00:42:27.380 align:middle line:84% and there are these repairs that you 00:42:27.380 --> 00:42:31.100 align:middle line:84% use to repair violations of them that you have created 00:42:31.100 --> 00:42:34.340 align:middle line:90% by adding suffixes to things. 00:42:34.340 --> 00:42:37.930 align:middle line:90% 00:42:37.930 --> 00:42:42.910 align:middle line:84% Yawelmani was the first example that people first-- 00:42:42.910 --> 00:42:46.990 align:middle line:84% made people think about phonological problems this way. 00:42:46.990 --> 00:42:50.290 align:middle line:84% It was described as a conspiracy, 00:42:50.290 --> 00:42:53.920 align:middle line:84% that is, the phonologists who first 00:42:53.920 --> 00:42:55.967 align:middle line:84% noticed this about Yawelmani were saying, 00:42:55.967 --> 00:42:58.300 align:middle line:84% yes, we have all these rules and they do various things. 00:42:58.300 --> 00:43:00.400 align:middle line:84% But they all have the consequence that you get rid 00:43:00.400 --> 00:43:02.140 align:middle line:90% of three consonant sequences. 00:43:02.140 --> 00:43:04.060 align:middle line:84% And so it's as though there's a conspiracy, 00:43:04.060 --> 00:43:06.130 align:middle line:84% and all of these different rules are conspiring 00:43:06.130 --> 00:43:08.200 align:middle line:84% to have the consequence that you never 00:43:08.200 --> 00:43:09.520 align:middle line:90% have three consonants in a row. 00:43:09.520 --> 00:43:10.975 align:middle line:84% And then we might want to know, why? 00:43:10.975 --> 00:43:13.183 align:middle line:84% Why doesn't it like having three consonants in a row? 00:43:13.183 --> 00:43:14.900 align:middle line:90% There's all kinds of work to do. 00:43:14.900 --> 00:43:18.400 align:middle line:84% I want to show you another example of a conspiracy. 00:43:18.400 --> 00:43:20.350 align:middle line:84% And we can probably get some ways into this 00:43:20.350 --> 00:43:22.060 align:middle line:90% before we have to stop. 00:43:22.060 --> 00:43:23.800 align:middle line:90% This one is from Japanese. 00:43:23.800 --> 00:43:25.360 align:middle line:84% So in order to show you this, I need 00:43:25.360 --> 00:43:26.410 align:middle line:84% to tell you some things about how 00:43:26.410 --> 00:43:27.580 align:middle line:90% Japanese pronunciation goes. 00:43:27.580 --> 00:43:32.350 align:middle line:84% This is all going to be from the Tokyo dialect of Japanese. 00:43:32.350 --> 00:43:36.630 align:middle line:84% So this is how pitch accent works in Japanese. 00:43:36.630 --> 00:43:40.260 align:middle line:90% So here are four words. 00:43:40.260 --> 00:43:42.120 align:middle line:84% Does anybody here speak Japanese natively? 00:43:42.120 --> 00:43:43.350 align:middle line:90% Any Japanese people here? 00:43:43.350 --> 00:43:44.970 align:middle line:90% Oh, cool. 00:43:44.970 --> 00:43:46.980 align:middle line:84% Can you pronounce these words for us? 00:43:46.980 --> 00:43:51.253 align:middle line:84% AUDIENCE: "Makura-wa, kokoro-wa, atama-wa, sakana-wa." 00:43:51.253 --> 00:43:52.170 align:middle line:90% NORVIN RICHARDS: Cool. 00:43:52.170 --> 00:43:56.520 align:middle line:84% So I just broke my heart listening to you 00:43:56.520 --> 00:43:57.780 align:middle line:90% say those words that way. 00:43:57.780 --> 00:43:59.370 align:middle line:84% There are various ways to pronounce these words. 00:43:59.370 --> 00:44:00.912 align:middle line:84% Let me pronounce them a different way 00:44:00.912 --> 00:44:03.810 align:middle line:84% and then you can tell me why I pronounced them wrong. 00:44:03.810 --> 00:44:07.560 align:middle line:84% The first one I think we agreed on, it's "makura-wa." 00:44:07.560 --> 00:44:09.960 align:middle line:84% There are speakers who will pronounce the second one 00:44:09.960 --> 00:44:14.850 align:middle line:84% "kokoro-wa," and the third one, "atama-wa," and the last one, 00:44:14.850 --> 00:44:16.640 align:middle line:90% "sakana-wa." 00:44:16.640 --> 00:44:20.090 align:middle line:84% But for you, that's apparently-- so you were pronouncing 00:44:20.090 --> 00:44:22.100 align:middle line:84% all but the first one the same way, 00:44:22.100 --> 00:44:25.400 align:middle line:84% I think, low on the first one, and high on all of the others. 00:44:25.400 --> 00:44:29.180 align:middle line:84% This is a point of variation between dialects of Japanese. 00:44:29.180 --> 00:44:31.250 align:middle line:84% So people from different parts of Japan, 00:44:31.250 --> 00:44:33.473 align:middle line:84% when I was living in Japan, I would talk 00:44:33.473 --> 00:44:34.640 align:middle line:90% about going to other places. 00:44:34.640 --> 00:44:37.880 align:middle line:84% And the people I was living in Tokyo, the people in Tokyo 00:44:37.880 --> 00:44:40.090 align:middle line:84% would say, oh, they speak Japanese 00:44:40.090 --> 00:44:41.090 align:middle line:90% so strangely over there. 00:44:41.090 --> 00:44:44.937 align:middle line:84% They don't say "HAshi," they say "haSHI." 00:44:44.937 --> 00:44:46.770 align:middle line:84% They put the high tone in a different place, 00:44:46.770 --> 00:44:48.270 align:middle line:84% put high pitch in a different place. 00:44:48.270 --> 00:44:52.790 align:middle line:90% 00:44:52.790 --> 00:44:54.380 align:middle line:84% So I want to show you some things 00:44:54.380 --> 00:44:56.540 align:middle line:84% about how accent works in Japanese, 00:44:56.540 --> 00:44:58.138 align:middle line:90% how accent is pronounced. 00:44:58.138 --> 00:44:59.930 align:middle line:84% So first, I'm going to tell you some things 00:44:59.930 --> 00:45:02.030 align:middle line:84% about the laws of accent realization, 00:45:02.030 --> 00:45:03.410 align:middle line:90% the way I was taught them. 00:45:03.410 --> 00:45:06.623 align:middle line:84% And then I'm curious, actually, where are you from, where? 00:45:06.623 --> 00:45:07.790 align:middle line:90% AUDIENCE: I'm half-Japanese. 00:45:07.790 --> 00:45:12.980 align:middle line:84% My mom's from Japan, so I just went to a Japanese school 00:45:12.980 --> 00:45:13.550 align:middle line:90% in the US. 00:45:13.550 --> 00:45:14.030 align:middle line:90% NORVIN RICHARDS: Oh, wow. 00:45:14.030 --> 00:45:14.420 align:middle line:90% Awesome. 00:45:14.420 --> 00:45:15.020 align:middle line:90% Cool. 00:45:15.020 --> 00:45:17.240 align:middle line:84% That's cool that you speak Japanese so well. 00:45:17.240 --> 00:45:20.930 align:middle line:84% My son, whose mother is Japanese, 00:45:20.930 --> 00:45:22.890 align:middle line:84% is attempting to follow in your footsteps. 00:45:22.890 --> 00:45:23.390 align:middle line:90% Yeah. 00:45:23.390 --> 00:45:25.920 align:middle line:84% I have to get you to tell me how it worked for you. 00:45:25.920 --> 00:45:26.420 align:middle line:90% Cool. 00:45:26.420 --> 00:45:27.878 align:middle line:84% So let me tell you something things 00:45:27.878 --> 00:45:31.700 align:middle line:84% about how accent realization works. 00:45:31.700 --> 00:45:34.310 align:middle line:84% By accent, we're talking about the pitch accent 00:45:34.310 --> 00:45:35.300 align:middle line:90% system of Japanese. 00:45:35.300 --> 00:45:37.142 align:middle line:84% So if you decide to study Japanese, 00:45:37.142 --> 00:45:38.600 align:middle line:84% one of the things you must learn is 00:45:38.600 --> 00:45:41.330 align:middle line:84% where the pitch of your voice needs to be high 00:45:41.330 --> 00:45:44.660 align:middle line:90% and where it needs to be low. 00:45:44.660 --> 00:45:48.960 align:middle line:84% And the basic rules for this in Tokyo Japanese go like this. 00:45:48.960 --> 00:45:53.930 align:middle line:84% There are words that don't have an accent anywhere. 00:45:53.930 --> 00:45:57.680 align:middle line:84% The example on the slide here is "fish," the last word. 00:45:57.680 --> 00:46:01.610 align:middle line:84% And in words like that, the first syllable is low 00:46:01.610 --> 00:46:02.990 align:middle line:90% and the others are high. 00:46:02.990 --> 00:46:06.652 align:middle line:84% So you get "saKANA-WA," low, high, high, high. 00:46:06.652 --> 00:46:09.550 align:middle line:90% 00:46:09.550 --> 00:46:12.490 align:middle line:84% And then, when there is accent, the accent 00:46:12.490 --> 00:46:16.510 align:middle line:84% perturbs this distribution of lows and highs 00:46:16.510 --> 00:46:20.810 align:middle line:84% in a particular way that I now tried to describe for you. 00:46:20.810 --> 00:46:22.870 align:middle line:90% It goes like this. 00:46:22.870 --> 00:46:26.080 align:middle line:84% The accented syllable is high, so "pillow" 00:46:26.080 --> 00:46:27.670 align:middle line:84% has accent on the first syllable. 00:46:27.670 --> 00:46:29.620 align:middle line:84% "Heart" has accent on the second syllable. 00:46:29.620 --> 00:46:32.260 align:middle line:84% "Head" has accent on the third syllable. 00:46:32.260 --> 00:46:35.170 align:middle line:84% So you get a high on the accented syllable, 00:46:35.170 --> 00:46:38.500 align:middle line:84% and every syllable after an accented syllable is low. 00:46:38.500 --> 00:46:42.100 align:middle line:84% So pillow is "MAkura-wa," high, low, low, low, 00:46:42.100 --> 00:46:45.160 align:middle line:84% because it's got accent on the first syllable. 00:46:45.160 --> 00:46:48.790 align:middle line:84% And then, heart and head have accent 00:46:48.790 --> 00:46:51.245 align:middle line:84% on places that are not the first syllable. 00:46:51.245 --> 00:46:52.870 align:middle line:84% And so they have high on their accented 00:46:52.870 --> 00:46:54.650 align:middle line:84% syllable and then low from then on. 00:46:54.650 --> 00:46:59.530 align:middle line:84% So heart is "koKOro-wa," low, high, low, low. 00:46:59.530 --> 00:47:01.270 align:middle line:84% Low on the first syllable because, well, 00:47:01.270 --> 00:47:03.370 align:middle line:84% low on the first syllable is the general thing, 00:47:03.370 --> 00:47:08.380 align:middle line:84% and high on the second syllable, and then low from then on. 00:47:08.380 --> 00:47:08.980 align:middle line:90% Don't worry. 00:47:08.980 --> 00:47:10.420 align:middle line:84% You're not going to be quizzed on this. 00:47:10.420 --> 00:47:12.170 align:middle line:84% Actually, I'm really just telling you this 00:47:12.170 --> 00:47:14.350 align:middle line:84% so that you will know why we might care about where 00:47:14.350 --> 00:47:15.540 align:middle line:90% accent goes in Japanese. 00:47:15.540 --> 00:47:18.040 align:middle line:84% Mostly what we're going to be talking about in just a second 00:47:18.040 --> 00:47:20.875 align:middle line:84% is what the rules are for where accent goes in Japanese. 00:47:20.875 --> 00:47:23.500 align:middle line:84% Though if none of this makes any sense, or if you're wondering, 00:47:23.500 --> 00:47:25.990 align:middle line:84% why on Earth are we talking about all these highs and lows? 00:47:25.990 --> 00:47:28.365 align:middle line:84% It's really just so you'll know what we're talking about. 00:47:28.365 --> 00:47:30.820 align:middle line:84% We're talking about what the rules are for what goes high 00:47:30.820 --> 00:47:33.190 align:middle line:90% and what goes low in Japanese. 00:47:33.190 --> 00:47:37.030 align:middle line:84% So there are these regular rules that go for "fish." 00:47:37.030 --> 00:47:38.980 align:middle line:84% It's like, if you have an unaccented word 00:47:38.980 --> 00:47:41.770 align:middle line:84% it's low at the beginning and high from then on. 00:47:41.770 --> 00:47:44.410 align:middle line:84% If you have any accents, the basic rule 00:47:44.410 --> 00:47:47.350 align:middle line:84% is, put a high on the accented syllable, low from then 00:47:47.350 --> 00:47:50.140 align:middle line:84% on, and otherwise, act like a fish. 00:47:50.140 --> 00:47:53.133 align:middle line:84% So "pillow" has a high tone at the beginning 00:47:53.133 --> 00:47:55.300 align:middle line:84% because it's accented on the first syllable and then 00:47:55.300 --> 00:47:57.010 align:middle line:84% it's low from then on, "MAkura-wa." 00:47:57.010 --> 00:48:00.010 align:middle line:84% And then, everything else tries to be low, high, high, high, 00:48:00.010 --> 00:48:01.612 align:middle line:84% with the overriding principle that you 00:48:01.612 --> 00:48:03.820 align:middle line:84% have to have a high on your accented syllable and low 00:48:03.820 --> 00:48:04.540 align:middle line:90% from then on. 00:48:04.540 --> 00:48:07.420 align:middle line:84% That's the basic way that this works. 00:48:07.420 --> 00:48:09.580 align:middle line:84% And I promised Raquel that when we got to Japanese 00:48:09.580 --> 00:48:13.690 align:middle line:84% I would tell you about a way in which languages sometimes 00:48:13.690 --> 00:48:17.320 align:middle line:84% distinguish verbs from nouns and other kinds of things. 00:48:17.320 --> 00:48:24.500 align:middle line:84% Japanese nouns, as you can see, get to have accent or not. 00:48:24.500 --> 00:48:28.090 align:middle line:84% So "fish" doesn't have an accent and the rest of these words 00:48:28.090 --> 00:48:28.930 align:middle line:90% have accents. 00:48:28.930 --> 00:48:31.120 align:middle line:84% And if they have an accent, the accent 00:48:31.120 --> 00:48:34.420 align:middle line:84% can be on any syllable of the word. 00:48:34.420 --> 00:48:38.110 align:middle line:84% Verbs in Japanese only get to choose whether they're accented 00:48:38.110 --> 00:48:38.710 align:middle line:90% or not. 00:48:38.710 --> 00:48:41.290 align:middle line:84% They don't get to choose where the accent is. 00:48:41.290 --> 00:48:43.163 align:middle line:84% And that's not uncommon for languages 00:48:43.163 --> 00:48:44.830 align:middle line:84% to make that kind of distinction to have 00:48:44.830 --> 00:48:48.520 align:middle line:84% a richer array of possible accent types in nouns 00:48:48.520 --> 00:48:50.510 align:middle line:90% than in verbs. 00:48:50.510 --> 00:48:53.860 align:middle line:84% That's one of the kinds of noun-verb phonological 00:48:53.860 --> 00:48:56.380 align:middle line:84% differences people sometimes settle on. 00:48:56.380 --> 00:48:58.713 align:middle line:84% So really, this is just here for your greater education. 00:48:58.713 --> 00:49:00.880 align:middle line:84% If any of you were thinking about learning Japanese, 00:49:00.880 --> 00:49:03.130 align:middle line:84% this is one of the things you will have to learn. 00:49:03.130 --> 00:49:04.948 align:middle line:84% I'll feel bad if, now, any of you 00:49:04.948 --> 00:49:06.490 align:middle line:84% were thinking about learning Japanese 00:49:06.490 --> 00:49:07.900 align:middle line:90% but have now changed your mind. 00:49:07.900 --> 00:49:09.220 align:middle line:90% It's not that bad, really. 00:49:09.220 --> 00:49:11.905 align:middle line:90% 00:49:11.905 --> 00:49:13.280 align:middle line:84% And then there are minimal pairs. 00:49:13.280 --> 00:49:17.060 align:middle line:84% So "rain" in Tokyo Japanese is "ame." 00:49:17.060 --> 00:49:21.230 align:middle line:84% "Candy" in Tokyo Japanese is "ame," 00:49:21.230 --> 00:49:23.810 align:middle line:84% so they have the same consonants and vowels in them 00:49:23.810 --> 00:49:26.750 align:middle line:84% in the same sequence, but different positions of pitch 00:49:26.750 --> 00:49:27.380 align:middle line:90% accent. 00:49:27.380 --> 00:49:29.630 align:middle line:84% "Candy" is unaccented and "rain" has an accent 00:49:29.630 --> 00:49:30.590 align:middle line:90% on the first syllable. 00:49:30.590 --> 00:49:32.300 align:middle line:84% Or the one that everybody always uses 00:49:32.300 --> 00:49:34.370 align:middle line:84% is "chopsticks," which is HAshi, "bridge," 00:49:34.370 --> 00:49:40.700 align:middle line:84% which is "haSHI," and "edge," which is "haSHI." 00:49:40.700 --> 00:49:42.170 align:middle line:84% "Bridge" and "edge," there's a lot 00:49:42.170 --> 00:49:44.420 align:middle line:84% of beautiful Japanese phonology trying to figure out 00:49:44.420 --> 00:49:46.712 align:middle line:84% whether there's a difference between the pronunciations 00:49:46.712 --> 00:49:47.712 align:middle line:90% of "bridge" and "edge." 00:49:47.712 --> 00:49:49.670 align:middle line:84% The place where everybody can hear a difference 00:49:49.670 --> 00:49:51.120 align:middle line:90% is when you add suffixes. 00:49:51.120 --> 00:49:55.310 align:middle line:84% So if you add, say, the nominative suffix, "-ga," 00:49:55.310 --> 00:49:59.840 align:middle line:90% to those, "edge" is like fish. 00:49:59.840 --> 00:50:03.840 align:middle line:84% So not only will edge be low, high, but then the "-ga" 00:50:03.840 --> 00:50:06.440 align:middle line:84% will also be high, so you get "haSHIGA." 00:50:06.440 --> 00:50:11.390 align:middle line:84% Whereas "bridge" has an accent, that means there'll be a high 00:50:11.390 --> 00:50:13.970 align:middle line:84% on the "-shi," but then it'll be low from then on. 00:50:13.970 --> 00:50:18.200 align:middle line:84% So "edge," the nominative is "haSHIGA," but "bridge," 00:50:18.200 --> 00:50:21.240 align:middle line:84% the nominative is "haSHIga," low, high, low. 00:50:21.240 --> 00:50:23.570 align:middle line:84% So all fun things to look forward 00:50:23.570 --> 00:50:26.270 align:middle line:84% to if you decide to try to learn Japanese. 00:50:26.270 --> 00:50:28.670 align:middle line:90% Now here's a generalization. 00:50:28.670 --> 00:50:32.180 align:middle line:84% And remember, I told you when we were doing Yawelmani 00:50:32.180 --> 00:50:34.370 align:middle line:84% and also when we were doing English, 00:50:34.370 --> 00:50:37.010 align:middle line:84% we're going to talk about various kinds of things 00:50:37.010 --> 00:50:39.470 align:middle line:84% this language does under different circumstances. 00:50:39.470 --> 00:50:42.560 align:middle line:84% But here's a generalization that holds in Japanese, 00:50:42.560 --> 00:50:43.880 align:middle line:90% at least Tokyo Japanese. 00:50:43.880 --> 00:50:46.910 align:middle line:90% Words have at most one accent. 00:50:46.910 --> 00:50:48.950 align:middle line:84% They don't have more than one accent. 00:50:48.950 --> 00:50:50.720 align:middle line:84% And so we're going to look at cases where 00:50:50.720 --> 00:50:53.570 align:middle line:84% words are going to experience a temptation 00:50:53.570 --> 00:50:55.220 align:middle line:90% to have more than one accent. 00:50:55.220 --> 00:50:56.990 align:middle line:84% And they're going to do various things 00:50:56.990 --> 00:51:00.495 align:middle line:84% to avoid that, various things, kind of like Yawelmani 00:51:00.495 --> 00:51:02.120 align:middle line:84% was experiencing the temptation to have 00:51:02.120 --> 00:51:05.450 align:middle line:84% three consonants in a row, and it saved itself in various ways 00:51:05.450 --> 00:51:07.005 align:middle line:90% under different circumstances. 00:51:07.005 --> 00:51:08.880 align:middle line:84% We're going to see the same thing in Japanese 00:51:08.880 --> 00:51:10.463 align:middle line:84% and the moral is going to be the same. 00:51:10.463 --> 00:51:13.250 align:middle line:84% Here's a language that has a bumper sticker. 00:51:13.250 --> 00:51:15.740 align:middle line:84% It has an inspirational poster on its wall. 00:51:15.740 --> 00:51:18.470 align:middle line:84% Don't have more than one accent in your words. 00:51:18.470 --> 00:51:21.680 align:middle line:84% There's a cute picture of a kitten or something under it. 00:51:21.680 --> 00:51:24.140 align:middle line:84% And then the way it achieves that is going 00:51:24.140 --> 00:51:25.340 align:middle line:90% to vary from word to word. 00:51:25.340 --> 00:51:28.490 align:middle line:84% That's going to be the moral of what we're going to see. 00:51:28.490 --> 00:51:29.450 align:middle line:90% Let's look at that. 00:51:29.450 --> 00:51:32.790 align:middle line:90% 00:51:32.790 --> 00:51:36.030 align:middle line:84% There's "pillow," "pillow" has accent on the first syllable. 00:51:36.030 --> 00:51:39.660 align:middle line:84% Yeah, so "from the pillow" is something like, "MAkura-kara," 00:51:39.660 --> 00:51:44.160 align:middle line:84% and "to the pillow" is something like, "MAkura-made." 00:51:44.160 --> 00:51:49.240 align:middle line:84% The heart, the head, yeah, same kinds of deal. 00:51:49.240 --> 00:51:52.140 align:middle line:84% So "from the heart" is "koKOro-kara," 00:51:52.140 --> 00:51:54.810 align:middle line:90% with accent just on "koKOro." 00:51:54.810 --> 00:51:58.215 align:middle line:84% To the heart is "koKoro-made," with accent just on the "-ko." 00:51:58.215 --> 00:52:01.980 align:middle line:90% 00:52:01.980 --> 00:52:04.020 align:middle line:84% "From the head" and "to the head," same deal. 00:52:04.020 --> 00:52:07.080 align:middle line:84% There's an accent on the last syllable of "head." 00:52:07.080 --> 00:52:09.300 align:middle line:84% But when we look at "fish," and remember 00:52:09.300 --> 00:52:11.370 align:middle line:84% that fish is underlyingly unaccented, 00:52:11.370 --> 00:52:13.980 align:middle line:84% that's why there aren't any accented vowels there 00:52:13.980 --> 00:52:16.230 align:middle line:84% in "sakana," the word for "fish." 00:52:16.230 --> 00:52:17.955 align:middle line:84% When we look at "fish," we look at it 00:52:17.955 --> 00:52:20.880 align:middle line:84% and we find a difference between "from" and "to." 00:52:20.880 --> 00:52:25.650 align:middle line:84% So for "from the fish," the whole sequence, "sakana-kara," 00:52:25.650 --> 00:52:27.330 align:middle line:90% is unaccented. 00:52:27.330 --> 00:52:31.350 align:middle line:84% But for "to the fish," the sequence "-made" 00:52:31.350 --> 00:52:32.670 align:middle line:90% has an accent on "-made." 00:52:32.670 --> 00:52:37.960 align:middle line:84% It's "sakana-MAde," with a dip at the end there. 00:52:37.960 --> 00:52:42.120 align:middle line:84% So "from" and "to" act the same for "pillow," "heart," 00:52:42.120 --> 00:52:43.950 align:middle line:90% and "head," the accented nouns. 00:52:43.950 --> 00:52:45.960 align:middle line:84% But for the unaccented noun, the word 00:52:45.960 --> 00:52:48.840 align:middle line:84% for "fish," which doesn't have an accent on its own, 00:52:48.840 --> 00:52:51.030 align:middle line:84% then we see a difference between "from" and "to." 00:52:51.030 --> 00:52:54.420 align:middle line:84% "From" doesn't have an accent but "to" acts like it 00:52:54.420 --> 00:52:55.740 align:middle line:90% has an accent all of a sudden. 00:52:55.740 --> 00:52:58.500 align:middle line:90% 00:52:58.500 --> 00:53:01.200 align:middle line:84% A conventional way of talking about what's going on 00:53:01.200 --> 00:53:07.050 align:middle line:84% says, yeah, there are, what, four, or five-- six! 00:53:07.050 --> 00:53:09.210 align:middle line:84% There are six morphemes in this slide. 00:53:09.210 --> 00:53:12.150 align:middle line:84% There's "pillow," "heart," "head," "fish," "from," and 00:53:12.150 --> 00:53:13.650 align:middle line:90% "to." 00:53:13.650 --> 00:53:15.960 align:middle line:84% "Pillow," "heart," and "head" are accented. 00:53:15.960 --> 00:53:17.520 align:middle line:90% "Fish" is unaccented. 00:53:17.520 --> 00:53:18.450 align:middle line:90% We already knew that. 00:53:18.450 --> 00:53:20.730 align:middle line:84% I showed you that on the earlier slide. 00:53:20.730 --> 00:53:22.470 align:middle line:90% "From" is also unaccented. 00:53:22.470 --> 00:53:27.180 align:middle line:84% "From" is like "fish," but "to" is accented, 00:53:27.180 --> 00:53:31.080 align:middle line:84% and that's why it's showing up as accented in that phrase that 00:53:31.080 --> 00:53:35.600 align:middle line:84% means "to the fish," "sakana-MAde." 00:53:35.600 --> 00:53:39.095 align:middle line:84% "Made" doesn't just mean "to," it means "as far as." 00:53:39.095 --> 00:53:41.960 align:middle line:84% It suggests that someone is crawling on the beach 00:53:41.960 --> 00:53:43.040 align:middle line:90% up to the fish. 00:53:43.040 --> 00:53:44.480 align:middle line:90% They got to the fish. 00:53:44.480 --> 00:53:49.030 align:middle line:84% They got that far, "sakana-MAde." 00:53:49.030 --> 00:53:55.480 align:middle line:84% So there are six morphemes on this slide and four 00:53:55.480 --> 00:54:00.340 align:middle line:84% of them are accented, "pillow," "heart," "head," and "to." 00:54:00.340 --> 00:54:03.610 align:middle line:84% So we only get to see that "to" is accented when we combine it 00:54:03.610 --> 00:54:07.390 align:middle line:84% with an unaccented noun, because there's this general principle 00:54:07.390 --> 00:54:11.060 align:middle line:84% that you can't have a word that has more than one accent in it. 00:54:11.060 --> 00:54:14.650 align:middle line:84% So when you add to to a noun that's accented, 00:54:14.650 --> 00:54:16.700 align:middle line:90% to loses its accent. 00:54:16.700 --> 00:54:18.090 align:middle line:84% "To" and "from" act the same way. 00:54:18.090 --> 00:54:19.040 align:middle line:90% Does that make sense? 00:54:19.040 --> 00:54:21.420 align:middle line:84% So here's case number one of Japanese working out 00:54:21.420 --> 00:54:23.170 align:middle line:84% a way to make sure that a single word only 00:54:23.170 --> 00:54:25.400 align:middle line:90% has a single accent in it. 00:54:25.400 --> 00:54:26.740 align:middle line:90% Here's a repair that it takes. 00:54:26.740 --> 00:54:30.690 align:middle line:84% It deletes the accent on "to," on the word for "to," "-made." 00:54:30.690 --> 00:54:32.440 align:middle line:84% There are other repairs you could imagine. 00:54:32.440 --> 00:54:35.690 align:middle line:84% It could have gotten rid of the accent on the noun instead, 00:54:35.690 --> 00:54:36.360 align:middle line:90% but it doesn't. 00:54:36.360 --> 00:54:37.360 align:middle line:90% That's not what it does. 00:54:37.360 --> 00:54:40.840 align:middle line:90% 00:54:40.840 --> 00:54:45.430 align:middle line:84% "Fish," "pillow," here's another Japanese word, "even," 00:54:45.430 --> 00:54:49.240 align:middle line:90% as in "even a fish." 00:54:49.240 --> 00:54:52.150 align:middle line:84% Try to imagine a circumstance in which you 00:54:52.150 --> 00:54:56.350 align:middle line:84% would want to say, "even a fish," or "even a pillow." 00:54:56.350 --> 00:54:58.480 align:middle line:90% "Even" is accented. 00:54:58.480 --> 00:55:01.870 align:middle line:84% We saw that "fish" is unaccented and "pillow" is accented. 00:55:01.870 --> 00:55:03.730 align:middle line:84% But when you add this word for even, 00:55:03.730 --> 00:55:08.140 align:middle line:84% "-gurai,' both of those nouns become unaccented. 00:55:08.140 --> 00:55:14.560 align:middle line:84% So it's sakana-GUrai, and it's makura-GUrai, even 00:55:14.560 --> 00:55:18.010 align:middle line:90% though it's MAkura and saKANA. 00:55:18.010 --> 00:55:21.770 align:middle line:84% So "pillow" is accented on its first syllable normally, 00:55:21.770 --> 00:55:25.230 align:middle line:84% but it loses its accent if you add this "even" thing. 00:55:25.230 --> 00:55:27.690 align:middle line:84% That's the repair that I just alluded to accidentally. 00:55:27.690 --> 00:55:31.470 align:middle line:84% So "even" is like "to" in that it has an accent. 00:55:31.470 --> 00:55:35.670 align:middle line:84% So if you were going to add "even" to these nouns, 00:55:35.670 --> 00:55:38.520 align:middle line:84% with "fish," no problem, with "pillow," there is a problem. 00:55:38.520 --> 00:55:40.380 align:middle line:84% You're in danger of having two accents. 00:55:40.380 --> 00:55:42.672 align:middle line:84% And the repair in this case is to get rid of the accent 00:55:42.672 --> 00:55:44.130 align:middle line:90% on "pillow." 00:55:44.130 --> 00:55:46.530 align:middle line:84% Raquel, did I bulldoze a question? 00:55:46.530 --> 00:55:49.235 align:middle line:84% AUDIENCE: I had a somewhat related question 00:55:49.235 --> 00:55:50.610 align:middle line:84% to the topic in general, which is 00:55:50.610 --> 00:55:54.040 align:middle line:84% just you're saying that languages will be like, I don't 00:55:54.040 --> 00:55:55.000 align:middle line:90% want this to happen? 00:55:55.000 --> 00:55:57.610 align:middle line:90% NORVIN RICHARDS: Yeah. 00:55:57.610 --> 00:55:59.440 align:middle line:84% AUDIENCE: Do we observe times in languages 00:55:59.440 --> 00:56:01.780 align:middle line:84% where they seem like they don't want things to happen, 00:56:01.780 --> 00:56:04.360 align:middle line:84% and they just break their rules all the time? 00:56:04.360 --> 00:56:06.850 align:middle line:84% Not like English and spelling, but something like they 00:56:06.850 --> 00:56:09.058 align:middle line:84% have certain things they definitely don't want to do, 00:56:09.058 --> 00:56:11.820 align:middle line:84% so you get just lots and lots of examples of times [INAUDIBLE]?? 00:56:11.820 --> 00:56:14.130 align:middle line:84% NORVIN RICHARDS: Of apparent counterexamples? 00:56:14.130 --> 00:56:17.070 align:middle line:84% Well, we saw something sort of like that 00:56:17.070 --> 00:56:20.400 align:middle line:90% when we were looking at Lardil. 00:56:20.400 --> 00:56:25.230 align:middle line:84% Maybe you remember, I showed you that it 00:56:25.230 --> 00:56:31.380 align:middle line:84% was useful to think that nouns don't 00:56:31.380 --> 00:56:34.920 align:middle line:90% like to end in the vowel "u." 00:56:34.920 --> 00:56:37.050 align:middle line:90% The vowel "u" changes to "a." 00:56:37.050 --> 00:56:40.390 align:middle line:84% So there was an underlying [NON-ENGLISH],, 00:56:40.390 --> 00:56:44.190 align:middle line:84% which means "blood," which you get to see in the accusative, 00:56:44.190 --> 00:56:45.490 align:middle line:90% which is [NON-ENGLISH]. 00:56:45.490 --> 00:56:48.310 align:middle line:84% But the nominative is [NON-ENGLISH].. 00:56:48.310 --> 00:56:50.700 align:middle line:84% And we said, well, that's because Lardil 00:56:50.700 --> 00:56:52.140 align:middle line:84% doesn't like words to end in "u." 00:56:52.140 --> 00:56:54.180 align:middle line:90% it changes the "u" to an "a." 00:56:54.180 --> 00:56:56.910 align:middle line:84% Lardil also doesn't like words to end in "k." 00:56:56.910 --> 00:56:59.130 align:middle line:84% So there's an underlying word for "boomerang," 00:56:59.130 --> 00:57:01.020 align:middle line:84% which is underlyingly [NON-ENGLISH],, 00:57:01.020 --> 00:57:02.430 align:middle line:90% which you see in the accusative. 00:57:02.430 --> 00:57:06.100 align:middle line:84% It's [NON-ENGLISH],, but the nominative is [NON-ENGLISH].. 00:57:06.100 --> 00:57:09.780 align:middle line:84% And so we said, yeah, it doesn't like words to end in "k." 00:57:09.780 --> 00:57:13.830 align:middle line:84% But Lardil also has words that are underlyingly 00:57:13.830 --> 00:57:15.810 align:middle line:90% ending in a "u" and then a "k." 00:57:15.810 --> 00:57:17.580 align:middle line:90% So this is the word for "story." 00:57:17.580 --> 00:57:21.630 align:middle line:84% The accusative is [NON-ENGLISH],, but the nominative is 00:57:21.630 --> 00:57:22.555 align:middle line:90% [NON-ENGLISH]. 00:57:22.555 --> 00:57:24.713 align:middle line:84% So you get rid of the "k" at the end, 00:57:24.713 --> 00:57:26.130 align:middle line:84% but now what you've got at the end 00:57:26.130 --> 00:57:29.070 align:middle line:84% is an "u," which we said when we looked at "blood," 00:57:29.070 --> 00:57:30.480 align:middle line:90% Lardil doesn't like. 00:57:30.480 --> 00:57:33.240 align:middle line:84% Is that the kind of thing you're talking about? 00:57:33.240 --> 00:57:33.930 align:middle line:90% AUDIENCE: Yeah. 00:57:33.930 --> 00:57:37.900 align:middle line:84% So just is it actually common that languages 00:57:37.900 --> 00:57:39.460 align:middle line:84% will break these important rules, 00:57:39.460 --> 00:57:42.535 align:middle line:84% or is it actually pretty infrequent? 00:57:42.535 --> 00:57:44.160 align:middle line:84% NORVIN RICHARDS: So we're going to see, 00:57:44.160 --> 00:57:45.940 align:middle line:84% I think it's fair to say that it's common. 00:57:45.940 --> 00:57:48.965 align:middle line:84% So I'm deliberately finding some cases, 00:57:48.965 --> 00:57:50.340 align:middle line:84% it may not feel this way, but I'm 00:57:50.340 --> 00:57:53.220 align:middle line:84% finding some cases that are comparatively simple, where 00:57:53.220 --> 00:57:58.120 align:middle line:84% there is a single principle that a language is trying to achieve 00:57:58.120 --> 00:57:59.790 align:middle line:84% and it achieves it in various ways. 00:57:59.790 --> 00:58:02.790 align:middle line:84% That's why we're talking about this kind of case. 00:58:02.790 --> 00:58:06.840 align:middle line:84% But there isn't any conflict between the various things 00:58:06.840 --> 00:58:08.590 align:middle line:90% that a language wants to do. 00:58:08.590 --> 00:58:11.310 align:middle line:84% So we're talking about cases where these languages only 00:58:11.310 --> 00:58:14.480 align:middle line:84% have one thing they want to do and they always achieve it. 00:58:14.480 --> 00:58:16.980 align:middle line:84% There are cases where a language has multiple things that it 00:58:16.980 --> 00:58:19.360 align:middle line:84% wants to do and they're not compatible with each other, 00:58:19.360 --> 00:58:20.820 align:middle line:90% and the language has to choose. 00:58:20.820 --> 00:58:22.440 align:middle line:84% So you can convince yourself that it 00:58:22.440 --> 00:58:24.130 align:middle line:84% wants to do both of these things, 00:58:24.130 --> 00:58:27.480 align:middle line:84% but sometimes it just has to give one of them up. 00:58:27.480 --> 00:58:29.460 align:middle line:84% And Lardil might be a case like that. 00:58:29.460 --> 00:58:31.150 align:middle line:84% That's a kind of thing that happens. 00:58:31.150 --> 00:58:33.600 align:middle line:84% So what we'll end up with is a picture in which languages 00:58:33.600 --> 00:58:36.060 align:middle line:84% might have many things that they want to do, 00:58:36.060 --> 00:58:40.170 align:middle line:84% but they have a list of priorities. 00:58:40.170 --> 00:58:42.600 align:middle line:84% And we're talking in these cases about languages that-- 00:58:42.600 --> 00:58:44.850 align:middle line:84% we're only talking about priority number one. 00:58:44.850 --> 00:58:45.870 align:middle line:90% It always wins. 00:58:45.870 --> 00:58:48.032 align:middle line:84% You always do whatever you have to to achieve that. 00:58:48.032 --> 00:58:50.490 align:middle line:84% But there are also languages that have priority number one, 00:58:50.490 --> 00:58:52.860 align:middle line:84% and number two, and number three, and number four. 00:58:52.860 --> 00:58:55.110 align:middle line:90% We find examples like that, too. 00:58:55.110 --> 00:58:56.565 align:middle line:90% Does that answer your question? 00:58:56.565 --> 00:58:57.370 align:middle line:90% AUDIENCE: Yeah. 00:58:57.370 --> 00:58:58.790 align:middle line:90% NORVIN RICHARDS: Cool. 00:58:58.790 --> 00:59:00.410 align:middle line:90% Yeah. 00:59:00.410 --> 00:59:01.946 align:middle line:90% Yeah? 00:59:01.946 --> 00:59:05.790 align:middle line:84% AUDIENCE: So taking the case to a study of English. 00:59:05.790 --> 00:59:09.825 align:middle line:84% I would assume that no double stridents 00:59:09.825 --> 00:59:10.912 align:middle line:90% is very high priority. 00:59:10.912 --> 00:59:11.787 align:middle line:90% NORVIN RICHARDS: Yes. 00:59:11.787 --> 00:59:13.070 align:middle line:84% AUDIENCE: Because I've never heard it. 00:59:13.070 --> 00:59:15.195 align:middle line:84% NORVIN RICHARDS: We don't have any words like that. 00:59:15.195 --> 00:59:16.283 align:middle line:90% Right. 00:59:16.283 --> 00:59:17.450 align:middle line:90% Yeah, that's a good example. 00:59:17.450 --> 00:59:19.550 align:middle line:84% So that's a case where English just 00:59:19.550 --> 00:59:22.460 align:middle line:84% will not tolerate cases where a word ends in two stridents. 00:59:22.460 --> 00:59:24.110 align:middle line:84% Yeah, we don't have words like that. 00:59:24.110 --> 00:59:25.850 align:middle line:90% Good example. 00:59:25.850 --> 00:59:27.150 align:middle line:90% Yeah. 00:59:27.150 --> 00:59:27.650 align:middle line:90% All right. 00:59:27.650 --> 00:59:33.120 align:middle line:84% Let me continue crunching through Japanese here. 00:59:33.120 --> 00:59:35.660 align:middle line:84% So again, Japanese doesn't allow for words 00:59:35.660 --> 00:59:36.890 align:middle line:90% with more than one accent. 00:59:36.890 --> 00:59:40.430 align:middle line:84% We've seen that "-made," this word for "to," loses its accent 00:59:40.430 --> 00:59:41.468 align:middle line:90% after an accented word. 00:59:41.468 --> 00:59:43.760 align:middle line:84% And now we're seeing that if you have an accented word, 00:59:43.760 --> 00:59:47.180 align:middle line:84% it will lose its accent before "-gurai." 00:59:47.180 --> 00:59:50.870 align:middle line:84% So again, Japanese has this general thing, 00:59:50.870 --> 00:59:53.400 align:middle line:84% very high priority, no words with more than one accent 00:59:53.400 --> 00:59:53.900 align:middle line:90% in them. 00:59:53.900 --> 00:59:55.687 align:middle line:84% And then it has various fixes depending 00:59:55.687 --> 00:59:57.020 align:middle line:90% on the particular circumstances. 00:59:57.020 --> 00:59:58.520 align:middle line:90% We've talked about two so far. 00:59:58.520 --> 01:00:01.370 align:middle line:90% We'll see some more. 01:00:01.370 --> 01:00:01.870 align:middle line:90% OK. 01:00:01.870 --> 01:00:06.250 align:middle line:84% So Japanese avoids having more than one accent in this unit 01:00:06.250 --> 01:00:08.590 align:middle line:84% consisting of a word along with its suffixes 01:00:08.590 --> 01:00:09.507 align:middle line:90% through various means. 01:00:09.507 --> 01:00:11.715 align:middle line:84% You either get rid of the first accent or the second, 01:00:11.715 --> 01:00:13.330 align:middle line:84% depending on which accent exactly it 01:00:13.330 --> 01:00:14.680 align:middle line:90% is that you're talking about. 01:00:14.680 --> 01:00:17.222 align:middle line:84% We can maybe think of this as an example of the kind of thing 01:00:17.222 --> 01:00:21.250 align:middle line:84% that [INAUDIBLE] Japanese's highest 01:00:21.250 --> 01:00:25.000 align:middle line:84% priority is, don't have a word with more than one accent. 01:00:25.000 --> 01:00:26.740 align:middle line:84% And then it has these other priorities, 01:00:26.740 --> 01:00:29.020 align:middle line:84% like keep the accent on the noun, 01:00:29.020 --> 01:00:33.070 align:middle line:84% and keep the accent on "-made," the word for "to," and keep 01:00:33.070 --> 01:00:35.680 align:middle line:84% the accent on "-gurai," the word for "even." 01:00:35.680 --> 01:00:38.230 align:middle line:84% And what we're seeing maybe is that we 01:00:38.230 --> 01:00:43.180 align:middle line:84% should list the priorities as only one accent. 01:00:43.180 --> 01:00:45.880 align:middle line:90% That's number one. 01:00:45.880 --> 01:00:49.990 align:middle line:84% And number two is keep the accent on "-gurai," 01:00:49.990 --> 01:00:51.550 align:middle line:90% this word that means "even." 01:00:51.550 --> 01:00:56.650 align:middle line:84% And number three is, keep the accent on nouns. 01:00:56.650 --> 01:01:04.150 align:middle line:84% And then, number four is, keep the accent on "-made," 01:01:04.150 --> 01:01:06.610 align:middle line:90% this word for "to." 01:01:06.610 --> 01:01:08.620 align:middle line:84% And these priorities we should think 01:01:08.620 --> 01:01:10.340 align:middle line:90% of as being in this order. 01:01:10.340 --> 01:01:12.220 align:middle line:84% So Japanese really only wants one accent 01:01:12.220 --> 01:01:14.840 align:middle line:84% per word, that's the most important thing. 01:01:14.840 --> 01:01:18.460 align:middle line:84% And then also, it really wants to keep the accent on "-gurai." 01:01:18.460 --> 01:01:20.650 align:middle line:84% It would like to keep the accent on nouns. 01:01:20.650 --> 01:01:23.240 align:middle line:84% So if it just has to choose between the accent on a noun 01:01:23.240 --> 01:01:25.490 align:middle line:84% and the accent on "-made," well, it'll keep the accent 01:01:25.490 --> 01:01:26.320 align:middle line:90% on the noun. 01:01:26.320 --> 01:01:28.403 align:middle line:90% It'll lose the accent on "to." 01:01:28.403 --> 01:01:31.070 align:middle line:84% But if it's choosing between the accent on a noun and the accent 01:01:31.070 --> 01:01:33.070 align:middle line:84% on "-gurai," on even, well, it'll choose the one 01:01:33.070 --> 01:01:34.600 align:middle line:90% on "-gurai." 01:01:34.600 --> 01:01:36.730 align:middle line:84% So it's as though we have, again, this 01:01:36.730 --> 01:01:38.650 align:middle line:84% ranked list of priorities, these things 01:01:38.650 --> 01:01:40.780 align:middle line:90% that Japanese cares about. 01:01:40.780 --> 01:01:43.480 align:middle line:84% And this sort of way of thinking about things 01:01:43.480 --> 01:01:45.910 align:middle line:84% raises all kinds of questions, like, what kinds of things 01:01:45.910 --> 01:01:47.980 align:middle line:84% do languages get to care about, and why? 01:01:47.980 --> 01:01:50.350 align:middle line:84% And what determines which things are more important 01:01:50.350 --> 01:01:51.350 align:middle line:90% than which other things? 01:01:51.350 --> 01:01:52.850 align:middle line:84% Is there anything to say about that? 01:01:52.850 --> 01:01:55.090 align:middle line:84% These are all open topics in phonology, 01:01:55.090 --> 01:01:56.620 align:middle line:90% things people work on. 01:01:56.620 --> 01:02:02.700 align:middle line:90% 01:02:02.700 --> 01:02:04.997 align:middle line:84% I'll show you a few more facts about Japanese 01:02:04.997 --> 01:02:05.955 align:middle line:90% before we have to stop. 01:02:05.955 --> 01:02:08.640 align:middle line:90% 01:02:08.640 --> 01:02:10.960 align:middle line:84% Here are some Japanese compounds. 01:02:10.960 --> 01:02:14.380 align:middle line:84% So compounds where you take two nouns and put them together. 01:02:14.380 --> 01:02:20.190 align:middle line:84% And of course, you may all be imagining potential problems, 01:02:20.190 --> 01:02:22.950 align:middle line:84% because as you know, a single word can only 01:02:22.950 --> 01:02:24.135 align:middle line:90% have one accent in it. 01:02:24.135 --> 01:02:27.330 align:middle line:90% 01:02:27.330 --> 01:02:31.260 align:middle line:84% What's striking about compounds is that the compound word, 01:02:31.260 --> 01:02:35.190 align:middle line:84% the word that you create when you put two words together, 01:02:35.190 --> 01:02:37.380 align:middle line:84% indeed, it can only have one accent in it. 01:02:37.380 --> 01:02:39.360 align:middle line:84% But the accent it has in it doesn't 01:02:39.360 --> 01:02:43.330 align:middle line:84% have to be the accent of either of the two components. 01:02:43.330 --> 01:02:46.330 align:middle line:84% So take the last example, there's a place. 01:02:46.330 --> 01:02:51.180 align:middle line:84% It's a place I lived in Japan for about a year, Chiba. 01:02:51.180 --> 01:02:51.900 align:middle line:90% Chiba ken. 01:02:51.900 --> 01:02:54.510 align:middle line:84% So Japan is divided into what are called prefectures. 01:02:54.510 --> 01:02:57.630 align:middle line:84% They're like states, so there are a certain number of them 01:02:57.630 --> 01:02:59.700 align:middle line:90% in Japan. 01:02:59.700 --> 01:03:01.800 align:middle line:84% There's a city that's the capital of Chiba 01:03:01.800 --> 01:03:03.780 align:middle line:90% prefecture, Chiba City. 01:03:03.780 --> 01:03:08.222 align:middle line:84% If you've ever read any William Gibson, you've read about it. 01:03:08.222 --> 01:03:09.180 align:middle line:90% That's science fiction. 01:03:09.180 --> 01:03:11.680 align:middle line:90% Chiba is not like that. 01:03:11.680 --> 01:03:15.630 align:middle line:84% The place itself has accent on the first syllable, 01:03:15.630 --> 01:03:17.790 align:middle line:90% so it's "CHIba." 01:03:17.790 --> 01:03:22.500 align:middle line:84% And the word for "prefecture" has an accent on it, 01:03:22.500 --> 01:03:23.640 align:middle line:90% it's "ken." 01:03:23.640 --> 01:03:27.598 align:middle line:84% But the compound, "Chiba prefecture," has one accent. 01:03:27.598 --> 01:03:29.640 align:middle line:84% That's sort of what you expect, words should only 01:03:29.640 --> 01:03:32.190 align:middle line:84% have one accent, but it is on the one syllable 01:03:32.190 --> 01:03:33.900 align:middle line:84% that it is not accented ordinarily. 01:03:33.900 --> 01:03:35.760 align:middle line:90% It's "chiBA ken." 01:03:35.760 --> 01:03:37.020 align:middle line:90% Sorry, "chiBA ken." 01:03:37.020 --> 01:03:39.120 align:middle line:84% That's right, "chiBA ken," so accent 01:03:39.120 --> 01:03:42.210 align:middle line:84% just on the second syllable, not the syllable 01:03:42.210 --> 01:03:44.005 align:middle line:84% that's ordinarily accented in "Chiba," 01:03:44.005 --> 01:03:46.380 align:middle line:84% and not the syllable that's ordinarily accented in "ken," 01:03:46.380 --> 01:03:47.070 align:middle line:90% the other one. 01:03:47.070 --> 01:03:49.770 align:middle line:90% 01:03:49.770 --> 01:03:51.570 align:middle line:84% You also get accent in compounds. 01:03:51.570 --> 01:03:53.730 align:middle line:84% You can see in the first example, 01:03:53.730 --> 01:03:57.570 align:middle line:84% you get accent in compounds even if the components 01:03:57.570 --> 01:04:00.390 align:middle line:84% of the compound are not accented themselves. 01:04:00.390 --> 01:04:04.410 align:middle line:84% So a "milk drinking child," "milk drinking" 01:04:04.410 --> 01:04:06.870 align:middle line:84% is itself a compound, but anyway, "milk drinking" 01:04:06.870 --> 01:04:07.680 align:middle line:90% is not accented. 01:04:07.680 --> 01:04:10.680 align:middle line:84% It's "tinomi," and "child" is not accented, it's "ko." 01:04:10.680 --> 01:04:14.880 align:middle line:84% But the compound has an accent, which is on the "i" there, 01:04:14.880 --> 01:04:17.640 align:middle line:90% so it's "tinoMIko." 01:04:17.640 --> 01:04:19.800 align:middle line:84% Yeah, so a nursing baby, "milk drinking child." 01:04:19.800 --> 01:04:22.540 align:middle line:90% 01:04:22.540 --> 01:04:24.580 align:middle line:84% Which prefecture is your mom from? 01:04:24.580 --> 01:04:25.495 align:middle line:90% Do you know? 01:04:25.495 --> 01:04:26.620 align:middle line:90% AUDIENCE: She's from Tokyo. 01:04:26.620 --> 01:04:28.037 align:middle line:84% NORVIN RICHARDS: She's from Tokyo? 01:04:28.037 --> 01:04:30.050 align:middle line:90% Cool. 01:04:30.050 --> 01:04:32.600 align:middle line:90% Great city, Tokyo. 01:04:32.600 --> 01:04:35.478 align:middle line:84% So yeah, so all of these compounds, 01:04:35.478 --> 01:04:37.520 align:middle line:84% they have one accent in them, exactly one accent. 01:04:37.520 --> 01:04:39.140 align:middle line:84% And the accent is not necessarily 01:04:39.140 --> 01:04:41.850 align:middle line:84% the accent of either of the components. 01:04:41.850 --> 01:04:43.520 align:middle line:84% In fact, I think I've carefully set it 01:04:43.520 --> 01:04:47.210 align:middle line:84% up so that it isn't in any of these examples. 01:04:47.210 --> 01:04:50.390 align:middle line:84% Where is the accent going in these compounds? 01:04:50.390 --> 01:04:51.590 align:middle line:90% Yes? 01:04:51.590 --> 01:04:53.190 align:middle line:84% AUDIENCE: The second to last syllable. 01:04:53.190 --> 01:04:55.023 align:middle line:84% NORVIN RICHARDS: So it's going on the second 01:04:55.023 --> 01:04:56.840 align:middle line:84% to last syllable in all of these examples. 01:04:56.840 --> 01:04:57.927 align:middle line:90% Yes. 01:04:57.927 --> 01:04:59.510 align:middle line:84% Is there another way you could say it? 01:04:59.510 --> 01:05:00.190 align:middle line:90% Yes? 01:05:00.190 --> 01:05:02.398 align:middle line:84% AUDIENCE: But would it be, like, on the last syllable 01:05:02.398 --> 01:05:04.050 align:middle line:84% of the first part of the compound? 01:05:04.050 --> 01:05:05.060 align:middle line:84% NORVIN RICHARDS: It's going on the last syllable 01:05:05.060 --> 01:05:06.530 align:middle line:84% of the first part of the compound. 01:05:06.530 --> 01:05:08.090 align:middle line:84% I haven't given you any examples here yet 01:05:08.090 --> 01:05:09.965 align:middle line:84% that distinguish those two theories but those 01:05:09.965 --> 01:05:13.125 align:middle line:84% are both perfectly good theories, what's going on. 01:05:13.125 --> 01:05:13.625 align:middle line:90% Yeah? 01:05:13.625 --> 01:05:17.690 align:middle line:90% 01:05:17.690 --> 01:05:21.930 align:middle line:84% Here are some places to distinguish them. 01:05:21.930 --> 01:05:23.880 align:middle line:90% I especially like the last one. 01:05:23.880 --> 01:05:28.230 align:middle line:84% So the Japanese word for "fried potato," or for french fries-- 01:05:28.230 --> 01:05:30.480 align:middle line:90% "furaido poteto," fried potato. 01:05:30.480 --> 01:05:36.540 align:middle line:90% 01:05:36.540 --> 01:05:38.430 align:middle line:84% Well, what we can see actually is 01:05:38.430 --> 01:05:42.210 align:middle line:84% that neither of those theories is right. 01:05:42.210 --> 01:05:45.870 align:middle line:84% If we ignore "nursing baby" and "Kagawa prefecture," 01:05:45.870 --> 01:05:49.710 align:middle line:84% where is accent going in "raw egg," and "field mouse," 01:05:49.710 --> 01:05:50.565 align:middle line:90% and "fried potato"? 01:05:50.565 --> 01:05:54.670 align:middle line:90% 01:05:54.670 --> 01:05:55.300 align:middle line:90% Yes? 01:05:55.300 --> 01:05:57.750 align:middle line:84% AUDIENCE: Isn't it called the antepenultimate? 01:05:57.750 --> 01:06:00.000 align:middle line:84% NORVIN RICHARDS: Oh, it is called the antepenultimate, 01:06:00.000 --> 01:06:01.500 align:middle line:90% yes. 01:06:01.500 --> 01:06:03.240 align:middle line:84% That is true in all of these examples. 01:06:03.240 --> 01:06:05.240 align:middle line:84% It's going on the antepenultimate syllable, that 01:06:05.240 --> 01:06:07.770 align:middle line:84% is, the syllable which is before the penultimate syllable, 01:06:07.770 --> 01:06:12.630 align:middle line:84% that is, the syllable which is three syllables from the end. 01:06:12.630 --> 01:06:14.675 align:middle line:84% Yeah, that's true for all of these. 01:06:14.675 --> 01:06:16.050 align:middle line:84% There's another way of saying it, 01:06:16.050 --> 01:06:19.720 align:middle line:84% though, sort of like there were two 01:06:19.720 --> 01:06:20.970 align:middle line:90% ways to say it the first time. 01:06:20.970 --> 01:06:21.330 align:middle line:90% Yeah? 01:06:21.330 --> 01:06:22.650 align:middle line:84% AUDIENCE: The first syllable of the second. 01:06:22.650 --> 01:06:23.220 align:middle line:84% NORVIN RICHARDS: The first syllable 01:06:23.220 --> 01:06:25.120 align:middle line:84% of the second part of the compound. 01:06:25.120 --> 01:06:27.372 align:middle line:84% Yeah, that works for all of these. 01:06:27.372 --> 01:06:29.580 align:middle line:84% I can't remember if I have any slides that show this, 01:06:29.580 --> 01:06:32.310 align:middle line:84% but that's actually the right way to talk about it. 01:06:32.310 --> 01:06:35.670 align:middle line:84% The right way to talk about where accent goes in a compound 01:06:35.670 --> 01:06:39.360 align:middle line:90% is, it goes at-- 01:06:39.360 --> 01:06:43.250 align:middle line:84% so first, it has to only have one accent. 01:06:43.250 --> 01:06:46.730 align:middle line:84% The accent has to be next to the boundary between the words 01:06:46.730 --> 01:06:50.360 align:middle line:84% and it can't be final, so it can't be on the last syllable. 01:06:50.360 --> 01:06:53.450 align:middle line:84% So in words like "Kagawa prefecture," 01:06:53.450 --> 01:06:56.300 align:middle line:84% so when you take "kagawa" and "ken," what you get 01:06:56.300 --> 01:06:57.800 align:middle line:90% is "kagawa ken." 01:06:57.800 --> 01:07:01.370 align:middle line:90% 01:07:01.370 --> 01:07:04.100 align:middle line:84% So there the accent goes at the end of "kagawa" 01:07:04.100 --> 01:07:06.183 align:middle line:90% because it can't be final. 01:07:06.183 --> 01:07:07.850 align:middle line:84% It can't be on the word for "prefecture" 01:07:07.850 --> 01:07:09.920 align:middle line:84% because that's only one syllable long. 01:07:09.920 --> 01:07:13.010 align:middle line:84% But on "fried potato," it can be on "potato," 01:07:13.010 --> 01:07:17.630 align:middle line:84% on the first syllable of "potato" 01:07:17.630 --> 01:07:19.790 align:middle line:84% because it's still between the boundary, 01:07:19.790 --> 01:07:22.400 align:middle line:84% next to the boundary between the words and it's not final. 01:07:22.400 --> 01:07:24.480 align:middle line:84% And then, if possible, that's what you prefer. 01:07:24.480 --> 01:07:30.110 align:middle line:84% So you prefer "furAIdo POteto" to "furaiDO poteto," 01:07:30.110 --> 01:07:34.280 align:middle line:84% where the accent would be at the end of the first syllable. 01:07:34.280 --> 01:07:35.930 align:middle line:90% Does that make sense? 01:07:35.930 --> 01:07:36.560 align:middle line:90% Make sense? 01:07:36.560 --> 01:07:40.050 align:middle line:90% 01:07:40.050 --> 01:07:45.000 align:middle line:90% So here's another place where-- 01:07:45.000 --> 01:07:48.870 align:middle line:84% so I did something like this over here. 01:07:48.870 --> 01:07:51.000 align:middle line:84% We'll just end with this and we'll pick it up here 01:07:51.000 --> 01:07:52.590 align:middle line:90% next time. 01:07:52.590 --> 01:07:55.860 align:middle line:84% Here's a place where it's useful to think of Japanese as having 01:07:55.860 --> 01:08:00.580 align:middle line:84% a bunch of priorities which it sort of has in an order. 01:08:00.580 --> 01:08:05.040 align:middle line:84% So if it's starting off with "kagawa" and "ken," 01:08:05.040 --> 01:08:09.880 align:middle line:84% Kagawa prefecture, it has to decide, what should I do? 01:08:09.880 --> 01:08:12.000 align:middle line:84% Should I keep both of the accents I've got? 01:08:12.000 --> 01:08:13.770 align:middle line:84% Should I just have one accent, and if I'm 01:08:13.770 --> 01:08:16.200 align:middle line:84% going to have one accent, where should it go? 01:08:16.200 --> 01:08:19.500 align:middle line:84% We've seen already from other examples in Japanese 01:08:19.500 --> 01:08:23.580 align:middle line:84% that it has a very high ranking preference for there only 01:08:23.580 --> 01:08:25.560 align:middle line:90% to be one accent. 01:08:25.560 --> 01:08:28.020 align:middle line:84% And so it's not going to keep two accents. 01:08:28.020 --> 01:08:30.901 align:middle line:84% That's the first candidate there, 01:08:30.901 --> 01:08:32.609 align:middle line:84% the first thing that it's imagining maybe 01:08:32.609 --> 01:08:34.350 align:middle line:90% being able to do. 01:08:34.350 --> 01:08:38.279 align:middle line:84% And then when it's trying to decide where the accent should 01:08:38.279 --> 01:08:42.420 align:middle line:84% be, the single accent, it has these various considerations. 01:08:42.420 --> 01:08:44.760 align:middle line:84% It wants the accent to be near the boundary. 01:08:44.760 --> 01:08:48.120 align:middle line:84% It would like the accent to be on the last word. 01:08:48.120 --> 01:08:50.939 align:middle line:84% But more important than having the accent on the last word 01:08:50.939 --> 01:08:53.819 align:middle line:84% is avoiding having accent on the last syllable. 01:08:53.819 --> 01:08:58.380 align:middle line:84% So when it's deciding between having 01:08:58.380 --> 01:09:00.149 align:middle line:84% the accent on the last syllable, which 01:09:00.149 --> 01:09:02.100 align:middle line:84% would put the accent on the last word, 01:09:02.100 --> 01:09:06.720 align:middle line:84% and putting the accent at the end of the first word 01:09:06.720 --> 01:09:08.984 align:middle line:84% in the compound, it chooses to put it 01:09:08.984 --> 01:09:10.859 align:middle line:84% at the end of the first word of the compound, 01:09:10.859 --> 01:09:13.420 align:middle line:84% because, well, that puts it near the boundary 01:09:13.420 --> 01:09:15.330 align:middle line:90% and avoids having it be final. 01:09:15.330 --> 01:09:19.710 align:middle line:84% So Japanese has these various things it's trying to do 01:09:19.710 --> 01:09:22.270 align:middle line:84% and it can't have everything that it wants. 01:09:22.270 --> 01:09:24.785 align:middle line:84% So it would like to accent the second word. 01:09:24.785 --> 01:09:26.910 align:middle line:84% It would like to put the accent on the second word. 01:09:26.910 --> 01:09:29.368 align:middle line:84% That's what we've seen when the second word is long enough, 01:09:29.368 --> 01:09:31.270 align:middle line:84% then the accent will go on the second word. 01:09:31.270 --> 01:09:34.034 align:middle line:84% But when the second word is so short 01:09:34.034 --> 01:09:36.630 align:middle line:84% that that would put it at the end of the compound word, 01:09:36.630 --> 01:09:40.859 align:middle line:84% then it doesn't go on the second word. 01:09:40.859 --> 01:09:43.000 align:middle line:84% It doesn't go at the end of the word. 01:09:43.000 --> 01:09:47.279 align:middle line:84% So avoiding having the accent at the end of the word 01:09:47.279 --> 01:09:52.450 align:middle line:84% is more important than putting the accent on the last word. 01:09:52.450 --> 01:09:53.939 align:middle line:84% So here's a place where Japanese is 01:09:53.939 --> 01:09:57.150 align:middle line:84% acting as though it has many things that it wants and it 01:09:57.150 --> 01:09:59.640 align:middle line:90% chooses the best one. 01:09:59.640 --> 01:10:01.920 align:middle line:84% This is a very influential way of thinking 01:10:01.920 --> 01:10:03.810 align:middle line:84% about phonological problems, which 01:10:03.810 --> 01:10:06.540 align:middle line:84% we will talk about more next time when we have more time. 01:10:06.540 --> 01:10:09.890 align:middle line:84% Are there any questions about it before I let you guys go? 01:10:09.890 --> 01:10:10.390 align:middle line:90% Yes? 01:10:10.390 --> 01:10:11.860 align:middle line:84% AUDIENCE: Does this have a formal name? 01:10:11.860 --> 01:10:12.735 align:middle line:90% NORVIN RICHARDS: Yes. 01:10:12.735 --> 01:10:14.852 align:middle line:84% It is called Optimality Theory, and we will 01:10:14.852 --> 01:10:16.060 align:middle line:90% talk more about it next time. 01:10:16.060 --> 01:10:19.000 align:middle line:84% You are seeking to make your choices optimal, 01:10:19.000 --> 01:10:23.460 align:middle line:84% to make them as good as they can be. 01:10:23.460 --> 01:10:24.890 align:middle line:90% Other questions? 01:10:24.890 --> 01:10:25.390 align:middle line:90% OK. 01:10:25.390 --> 01:10:28.750 align:middle line:84% So again, I'll put up a new problem set today, 01:10:28.750 --> 01:10:31.770 align:middle line:90% and I'll see you guys next time. 01:10:31.770 --> 01:10:39.000 align:middle line:90%