WEBVTT 00:00:00.000 --> 00:00:11.410 align:middle line:90% 00:00:11.410 --> 00:00:13.810 align:middle line:84% NORVIN RICHARDS: All right, now my mic is on. 00:00:13.810 --> 00:00:15.350 align:middle line:90% That has an effect on my voice. 00:00:15.350 --> 00:00:17.350 align:middle line:90% I'm being amplified, right? 00:00:17.350 --> 00:00:19.030 align:middle line:90% Cool. 00:00:19.030 --> 00:00:21.460 align:middle line:84% So it'll be easy for you to tell whether I've 00:00:21.460 --> 00:00:24.010 align:middle line:90% got my mic on or not. 00:00:24.010 --> 00:00:25.630 align:middle line:84% So here's another assimilation rule. 00:00:25.630 --> 00:00:28.270 align:middle line:84% We said, in Arabic, there's an assimilation rule-- wow, 00:00:28.270 --> 00:00:29.980 align:middle line:90% I'm really being amplified-- 00:00:29.980 --> 00:00:31.780 align:middle line:84% in Arabic, there's an assimilation rule 00:00:31.780 --> 00:00:34.420 align:middle line:84% that creates or causes two consonants 00:00:34.420 --> 00:00:35.780 align:middle line:90% to be identical to each other. 00:00:35.780 --> 00:00:36.940 align:middle line:90% This one is not that. 00:00:36.940 --> 00:00:42.400 align:middle line:84% The consonant at the end of that prefix is always nasal. 00:00:42.400 --> 00:00:46.010 align:middle line:84% It just assimilates in place to the following consonants. 00:00:46.010 --> 00:00:48.730 align:middle line:84% You get "impossible" and "incredible," 00:00:48.730 --> 00:00:50.762 align:middle line:84% where you get a velar nasal version of "n." 00:00:50.762 --> 00:00:52.720 align:middle line:84% It's not "incredible" [with alveolar nasal "n"] 00:00:52.720 --> 00:00:55.420 align:middle line:84% at least, unless you're speaking very carefully. 00:00:55.420 --> 00:00:56.535 align:middle line:90% That sound right? 00:00:56.535 --> 00:01:06.420 align:middle line:90% 00:01:06.420 --> 00:01:08.880 align:middle line:84% Well, we're talking about assimilation rules. 00:01:08.880 --> 00:01:11.818 align:middle line:84% I want to show you another example of vowel harmony. 00:01:11.818 --> 00:01:13.860 align:middle line:84% We looked at one example of harmony before, which 00:01:13.860 --> 00:01:16.110 align:middle line:90% was from Turkish. 00:01:16.110 --> 00:01:21.840 align:middle line:84% And it was the observation that the Turkish plural suffix is 00:01:21.840 --> 00:01:25.470 align:middle line:84% sometimes "-lar" and sometimes "-ler," 00:01:25.470 --> 00:01:28.980 align:middle line:84% and it depends on the vowels that are in the noun 00:01:28.980 --> 00:01:30.210 align:middle line:90% that's being made plural. 00:01:30.210 --> 00:01:34.080 align:middle line:84% So "lions" was "aslanlar"-- which I cannot spell all 00:01:34.080 --> 00:01:36.480 align:middle line:90% of a sudden. 00:01:36.480 --> 00:01:42.598 align:middle line:84% But "roses," I think, was "gyller." 00:01:42.598 --> 00:01:44.640 align:middle line:84% It's either "roses" or "flowers," I forget which. 00:01:44.640 --> 00:01:46.020 align:middle line:90% Oh, we have a native speaker. 00:01:46.020 --> 00:01:48.960 align:middle line:90% What does this mean, "gyller"? 00:01:48.960 --> 00:01:50.220 align:middle line:90% Is it "rose" or-- 00:01:50.220 --> 00:01:53.670 align:middle line:90% "rose," OK, good. 00:01:53.670 --> 00:01:56.250 align:middle line:84% So yeah, "aslanlar" but "gyller"-- why? 00:01:56.250 --> 00:01:58.710 align:middle line:84% Well, because this, we were going to say, 00:01:58.710 --> 00:02:01.830 align:middle line:84% is harmonizing with the vowels that 00:02:01.830 --> 00:02:04.180 align:middle line:90% are in the noun in question. 00:02:04.180 --> 00:02:07.110 align:middle line:84% So in particular, it's deciding whether to be 00:02:07.110 --> 00:02:11.310 align:middle line:84% back or front depending on whether the vowels in the noun 00:02:11.310 --> 00:02:12.198 align:middle line:90% are back or front. 00:02:12.198 --> 00:02:14.490 align:middle line:84% So I want to show you another example of vowel harmony, 00:02:14.490 --> 00:02:16.350 align:middle line:84% and we'll talk about it a little bit more. 00:02:16.350 --> 00:02:18.030 align:middle line:84% But that's what vowel harmony was. 00:02:18.030 --> 00:02:20.430 align:middle line:84% And as you can see, it's a kind of assimilation rule. 00:02:20.430 --> 00:02:24.750 align:middle line:84% So here's a vowel, and it's deciding whether to be-- it's 00:02:24.750 --> 00:02:28.650 align:middle line:84% gaining its front or back property for-- it's 00:02:28.650 --> 00:02:33.210 align:middle line:84% becoming the same as the vowels that are in the noun. 00:02:33.210 --> 00:02:35.250 align:middle line:84% So let's talk about another vowel harmony case. 00:02:35.250 --> 00:02:37.500 align:middle line:90% This is from Finnish. 00:02:37.500 --> 00:02:39.780 align:middle line:90% So here are the Finnish vowels. 00:02:39.780 --> 00:02:41.400 align:middle line:90% You can see it has-- 00:02:41.400 --> 00:02:43.500 align:middle line:84% what does it have-- eight of them, 00:02:43.500 --> 00:02:46.200 align:middle line:84% and they can all be long or short, but long or short 00:02:46.200 --> 00:02:48.690 align:middle line:84% for this just means how long you hold them. 00:02:48.690 --> 00:02:51.023 align:middle line:90% 00:02:51.023 --> 00:02:52.440 align:middle line:84% Since we're talking about Finnish, 00:02:52.440 --> 00:03:01.590 align:middle line:84% I have to tell you my favorite Finnish word, which is this, 00:03:01.590 --> 00:03:04.380 align:middle line:90% [NON-ENGLISH SPEAKING]. 00:03:04.380 --> 00:03:06.750 align:middle line:84% I was taught that word by a linguist 00:03:06.750 --> 00:03:08.070 align:middle line:90% who specialized in Finnish-- 00:03:08.070 --> 00:03:11.313 align:middle line:84% Erica Mitchell-- It's the only Finnish word I know. 00:03:11.313 --> 00:03:13.230 align:middle line:84% [NON-ENGLISH SPEAKING]---- I figure it'll come 00:03:13.230 --> 00:03:17.070 align:middle line:84% in conversation at some point, if I ever get back to Finland. 00:03:17.070 --> 00:03:19.308 align:middle line:84% It means wedding night intention. 00:03:19.308 --> 00:03:21.600 align:middle line:84% So I'm not sure what kind of conversation that will be, 00:03:21.600 --> 00:03:23.830 align:middle line:90% but I have to work on that. 00:03:23.830 --> 00:03:25.900 align:middle line:84% So there's the Finnish vowel inventory. 00:03:25.900 --> 00:03:28.680 align:middle line:84% One of the things you can see about this inventory-- 00:03:28.680 --> 00:03:31.860 align:middle line:84% it's got front vowels and back vowels. 00:03:31.860 --> 00:03:34.510 align:middle line:84% The vowels can be high or mid or low. 00:03:34.510 --> 00:03:39.930 align:middle line:84% And it is like English in that its back vowels are 00:03:39.930 --> 00:03:42.090 align:middle line:90% rounded if they are not low. 00:03:42.090 --> 00:03:46.560 align:middle line:84% So the back vowels are "u," "o," and "a." 00:03:46.560 --> 00:03:51.000 align:middle line:84% And so it has a high back rounded vowel-- for example, 00:03:51.000 --> 00:03:51.780 align:middle line:90% "u." 00:03:51.780 --> 00:03:56.190 align:middle line:84% But it doesn't have a high back unrounded vowel, "ouh." 00:03:56.190 --> 00:03:57.960 align:middle line:90% So far so much like English. 00:03:57.960 --> 00:04:00.780 align:middle line:84% Difference from English is that it also 00:04:00.780 --> 00:04:02.820 align:middle line:90% has front rounded vowels. 00:04:02.820 --> 00:04:05.220 align:middle line:84% So front high rounded vowel, front mid rounded vowel, 00:04:05.220 --> 00:04:07.590 align:middle line:90% "eouh" and "ouh." 00:04:07.590 --> 00:04:10.740 align:middle line:84% So English doesn't have those, but Finnish does. 00:04:10.740 --> 00:04:13.980 align:middle line:84% On the other hand, it also has front high and mid 00:04:13.980 --> 00:04:16.110 align:middle line:90% unrounded vowels, "i" and "e"-- 00:04:16.110 --> 00:04:17.760 align:middle line:90% so the pairs of vowels there. 00:04:17.760 --> 00:04:20.279 align:middle line:84% So I'm going to erase my favorite Finnish word, just so 00:04:20.279 --> 00:04:22.275 align:middle line:90% as not to confuse later classes. 00:04:22.275 --> 00:04:25.780 align:middle line:90% 00:04:25.780 --> 00:04:27.700 align:middle line:84% OK, so that's the Finnish vowel inventory. 00:04:27.700 --> 00:04:31.570 align:middle line:84% Now Finnish has vowel harmony like Turkish. 00:04:31.570 --> 00:04:39.490 align:middle line:84% It has a lot of suffixes that harmonize with the noun 00:04:39.490 --> 00:04:42.550 align:middle line:84% and, in particular, Finnish is kind of famous among linguists 00:04:42.550 --> 00:04:45.520 align:middle line:84% for having a lot of suffixes on nouns 00:04:45.520 --> 00:04:47.390 align:middle line:90% that are thought of as cases. 00:04:47.390 --> 00:04:51.580 align:middle line:84% So if you study languages like Latin or Russian 00:04:51.580 --> 00:04:53.905 align:middle line:84% or I bet Ukrainian, you need to learn 00:04:53.905 --> 00:04:55.780 align:middle line:84% that nouns come in a bunch of different forms 00:04:55.780 --> 00:04:57.238 align:middle line:84% depending on how they're being used 00:04:57.238 --> 00:04:59.507 align:middle line:84% in the sentence, the nominative or accusative-- 00:04:59.507 --> 00:05:01.090 align:middle line:84% they get little markers on them saying 00:05:01.090 --> 00:05:03.580 align:middle line:84% whether the subject or the object or the indirect object 00:05:03.580 --> 00:05:04.490 align:middle line:90% or whatever. 00:05:04.490 --> 00:05:06.430 align:middle line:84% This is what we call the case of a noun. 00:05:06.430 --> 00:05:09.370 align:middle line:84% Finnish-- languages that have case, 00:05:09.370 --> 00:05:13.120 align:middle line:84% they often have two cases or three or five or six. 00:05:13.120 --> 00:05:14.900 align:middle line:84% Finnish, depending on how you count, 00:05:14.900 --> 00:05:16.750 align:middle line:90% it has something like 12 or 13. 00:05:16.750 --> 00:05:18.460 align:middle line:90% It's just nuts. 00:05:18.460 --> 00:05:21.580 align:middle line:84% And that's partly because it has a bunch of case suffixes 00:05:21.580 --> 00:05:24.670 align:middle line:84% that, in English, would be prepositions. 00:05:24.670 --> 00:05:27.040 align:middle line:84% So it has a case suffix that means "on." 00:05:27.040 --> 00:05:29.170 align:middle line:84% This is the case suffix that means "on." 00:05:29.170 --> 00:05:30.685 align:middle line:84% And it is either pronounced "laa" 00:05:30.685 --> 00:05:32.977 align:middle line:84% [WITH LOW FRONT VOWEL] or "lah" [WITH LOW BACK VOWEL],, 00:05:32.977 --> 00:05:34.210 align:middle line:90% depending on the noun. 00:05:34.210 --> 00:05:36.790 align:middle line:84% So the word for "table" is [NON-ENGLISH SPEAKING] 00:05:36.790 --> 00:05:39.032 align:middle line:84% and "on the table" is [NON-ENGLISH SPEAKING].. 00:05:39.032 --> 00:05:40.990 align:middle line:84% The word for "street" is [NON-ENGLISH SPEAKING] 00:05:40.990 --> 00:05:45.730 align:middle line:84% and "on the street" is called [NON-ENGLISH SPEAKING].. 00:05:45.730 --> 00:05:48.010 align:middle line:84% I should be careful saying things like the word 00:05:48.010 --> 00:05:48.760 align:middle line:84% for "street" is [NON-ENGLISH SPEAKING].. 00:05:48.760 --> 00:05:50.700 align:middle line:84% I think the word for "street" is actually [NON-ENGLISH SPEAKING] 00:05:50.700 --> 00:05:51.828 align:middle line:90% if you have it by itself. 00:05:51.828 --> 00:05:53.620 align:middle line:84% But if you add the [NON-ENGLISH SPEAKING],, 00:05:53.620 --> 00:05:55.450 align:middle line:84% it changes the "t" to a "d," something like that. 00:05:55.450 --> 00:05:56.560 align:middle line:84% There are all kinds of sneaky things 00:05:56.560 --> 00:05:58.850 align:middle line:84% that happen in Finnish, that I forget, when you add them. 00:05:58.850 --> 00:06:00.725 align:middle line:84% But anyway, if you add [NON-ENGLISH SPEAKING] 00:06:00.725 --> 00:06:03.010 align:middle line:84% to street, you end up with [NON-ENGLISH SPEAKING].. 00:06:03.010 --> 00:06:04.385 align:middle line:84% If you add [NON-ENGLISH SPEAKING] 00:06:04.385 --> 00:06:06.160 align:middle line:84% to table, you get [NON-ENGLISH SPEAKING],, 00:06:06.160 --> 00:06:12.130 align:middle line:84% and the change in the vowel is conditioned by the vowel 00:06:12.130 --> 00:06:14.920 align:middle line:90% of the noun. 00:06:14.920 --> 00:06:16.870 align:middle line:84% Now I've kind of color-coded the vowels here. 00:06:16.870 --> 00:06:20.350 align:middle line:84% You can see there are some front vowels that are red 00:06:20.350 --> 00:06:22.610 align:middle line:84% and some back vowels that are blue. 00:06:22.610 --> 00:06:26.500 align:middle line:84% And if you have a noun where all the vowels are red, 00:06:26.500 --> 00:06:29.290 align:middle line:84% like "table," then the vowel of the suffix is also red. 00:06:29.290 --> 00:06:32.800 align:middle line:84% You have a vowel where all the-- sorry, a word where 00:06:32.800 --> 00:06:36.440 align:middle line:84% all the vowels are blue, then the suffix is also blue. 00:06:36.440 --> 00:06:40.780 align:middle line:84% So if all the vowels in the noun are back, 00:06:40.780 --> 00:06:43.570 align:middle line:84% like in "the streets," then the suffix gets a back vowel. 00:06:43.570 --> 00:06:47.350 align:middle line:84% If they're all front, then the suffix gets a front vowel. 00:06:47.350 --> 00:06:48.070 align:middle line:90% So far so good? 00:06:48.070 --> 00:06:50.630 align:middle line:90% 00:06:50.630 --> 00:06:53.700 align:middle line:90% Now here's the thing. 00:06:53.700 --> 00:06:55.500 align:middle line:90% There are also the green vowels. 00:06:55.500 --> 00:06:58.740 align:middle line:90% Why do I have green vowels? 00:06:58.740 --> 00:07:03.000 align:middle line:84% The red vowels all have the property that there is a blue 00:07:03.000 --> 00:07:05.520 align:middle line:90% vowel-- so a back vowel-- 00:07:05.520 --> 00:07:09.030 align:middle line:84% that is identical to the red vowel in every way 00:07:09.030 --> 00:07:11.490 align:middle line:84% except that it's back instead of front. 00:07:11.490 --> 00:07:14.680 align:middle line:84% So the red vowel "eouh" for example, 00:07:14.680 --> 00:07:18.630 align:middle line:84% the front high rounded vowel, has a corresponding blue vowel 00:07:18.630 --> 00:07:22.620 align:middle line:84% that is high and rounded and different only 00:07:22.620 --> 00:07:23.520 align:middle line:90% in that it's back. 00:07:23.520 --> 00:07:25.500 align:middle line:90% That's "u." 00:07:25.500 --> 00:07:29.670 align:middle line:84% And similarly, there is a front mid rounded vowel, 00:07:29.670 --> 00:07:32.190 align:middle line:84% "eouh," and there is another vowel 00:07:32.190 --> 00:07:33.690 align:middle line:84% that is identical to it in every way 00:07:33.690 --> 00:07:35.565 align:middle line:84% except that it's back instead of being front. 00:07:35.565 --> 00:07:38.190 align:middle line:90% That's oh. 00:07:38.190 --> 00:07:42.090 align:middle line:84% But the green vowels don't have any such correspondence. 00:07:42.090 --> 00:07:45.810 align:middle line:84% So "i," the front high unrounded vowel, 00:07:45.810 --> 00:07:48.000 align:middle line:84% there is no vowel which is identical to that vowel 00:07:48.000 --> 00:07:50.430 align:middle line:84% in every way except that it's back. 00:07:50.430 --> 00:07:52.470 align:middle line:90% There isn't an "ooh." 00:07:52.470 --> 00:07:54.460 align:middle line:90% Does that make sense? 00:07:54.460 --> 00:07:57.090 align:middle line:84% So the red vowels and the blue vowels 00:07:57.090 --> 00:07:59.850 align:middle line:84% are identical in every way except for front versus back. 00:07:59.850 --> 00:08:03.420 align:middle line:84% The green vowels don't have a corresponding back vowel 00:08:03.420 --> 00:08:05.970 align:middle line:84% that's identical to them except being back. 00:08:05.970 --> 00:08:06.870 align:middle line:90% Yeah? 00:08:06.870 --> 00:08:08.602 align:middle line:90% AUDIENCE: [INAUDIBLE] 00:08:08.602 --> 00:08:10.560 align:middle line:84% NORVIN RICHARDS: Well, it would be-- let's see, 00:08:10.560 --> 00:08:12.930 align:middle line:90% it would be "oueh." 00:08:12.930 --> 00:08:14.490 align:middle line:84% It's not a popular vowel, but there 00:08:14.490 --> 00:08:17.940 align:middle line:90% are languages that have it. 00:08:17.940 --> 00:08:22.770 align:middle line:84% The IPA symbol for it, I think, is something like this. 00:08:22.770 --> 00:08:24.540 align:middle line:90% So it exists. 00:08:24.540 --> 00:08:26.640 align:middle line:84% We haven't talked about a language that has it. 00:08:26.640 --> 00:08:29.150 align:middle line:90% 00:08:29.150 --> 00:08:32.331 align:middle line:84% I'll erase it quickly before it upsets anybody. 00:08:32.331 --> 00:08:33.789 align:middle line:84% So it's physically possible to make 00:08:33.789 --> 00:08:35.747 align:middle line:84% these vowels in their languages that have them, 00:08:35.747 --> 00:08:37.100 align:middle line:90% but Finnish doesn't have them. 00:08:37.100 --> 00:08:38.558 align:middle line:84% And so the question is, what do you 00:08:38.558 --> 00:08:42.760 align:middle line:84% do when you add a suffix to a noun that has these vowels? 00:08:42.760 --> 00:08:45.790 align:middle line:84% And the answer is that these vessels are just invisible, 00:08:45.790 --> 00:08:46.630 align:middle line:90% basically. 00:08:46.630 --> 00:08:51.250 align:middle line:84% So at the station is the less interesting case-- 00:08:51.250 --> 00:08:53.200 align:middle line:84% you have a couple of front vowels 00:08:53.200 --> 00:08:55.933 align:middle line:84% that do have corresponding back vowels, "y" and "ae," 00:08:55.933 --> 00:08:57.850 align:middle line:84% and then you have this "i," which doesn't have 00:08:57.850 --> 00:08:58.980 align:middle line:90% a corresponding back vowel. 00:08:58.980 --> 00:09:00.730 align:middle line:84% You have two red vowels and a green vowel. 00:09:00.730 --> 00:09:02.240 align:middle line:84% And you get a red vowel in the suffix, 00:09:02.240 --> 00:09:03.615 align:middle line:84% so you get [NON-ENGLISH SPEAKING] 00:09:03.615 --> 00:09:06.210 align:middle line:90% for at the station. 00:09:06.210 --> 00:09:10.610 align:middle line:84% But for "on the child," [NON-ENGLISH SPEAKING],, well, 00:09:10.610 --> 00:09:14.195 align:middle line:84% you have a blue vowel that is a back low vowel, 00:09:14.195 --> 00:09:17.750 align:middle line:84% "a," and then you have a green vowel, which is front, 00:09:17.750 --> 00:09:20.810 align:middle line:84% but the green vowel doesn't have a corresponding back vowel, 00:09:20.810 --> 00:09:22.670 align:middle line:84% and so it's like it doesn't count. 00:09:22.670 --> 00:09:25.080 align:middle line:84% The green vowels can go with anything. 00:09:25.080 --> 00:09:27.470 align:middle line:84% So you have vowel harmony for the vowels that 00:09:27.470 --> 00:09:30.153 align:middle line:84% have corresponding front and back versions, 00:09:30.153 --> 00:09:32.570 align:middle line:84% but for those that don't have corresponding front and back 00:09:32.570 --> 00:09:35.030 align:middle line:84% versions, it's like you get to ignore them. 00:09:35.030 --> 00:09:37.580 align:middle line:90% They're not there. 00:09:37.580 --> 00:09:38.200 align:middle line:90% Is that clear? 00:09:38.200 --> 00:09:39.030 align:middle line:90% That makes sense? 00:09:39.030 --> 00:09:40.070 align:middle line:90% So there are-- 00:09:40.070 --> 00:09:44.720 align:middle line:84% I mean, it's as though you can say as long as you only 00:09:44.720 --> 00:09:47.780 align:middle line:84% look at blue and red vowels, all the vowels in a word 00:09:47.780 --> 00:09:49.700 align:middle line:90% must be blue or red. 00:09:49.700 --> 00:09:51.722 align:middle line:84% But you don't look at green vowels. 00:09:51.722 --> 00:09:53.180 align:middle line:84% You don't look at vowels that don't 00:09:53.180 --> 00:09:55.970 align:middle line:84% have a corresponding back vowel that is identical in every way 00:09:55.970 --> 00:09:58.445 align:middle line:90% except for being back. 00:09:58.445 --> 00:09:59.820 align:middle line:84% So you get [NON-ENGLISH SPEAKING] 00:09:59.820 --> 00:10:02.940 align:middle line:90% for "on the child." 00:10:02.940 --> 00:10:06.570 align:middle line:84% There are also words in which all the vowels are green, 00:10:06.570 --> 00:10:10.410 align:middle line:84% like "at the family," and the fact 00:10:10.410 --> 00:10:14.070 align:middle line:84% is that for suffixes, you get front vowels corresponding 00:10:14.070 --> 00:10:14.640 align:middle line:90% to those. 00:10:14.640 --> 00:10:17.070 align:middle line:84% And the green vowels are front, which 00:10:17.070 --> 00:10:18.400 align:middle line:90% is itself kind of interesting. 00:10:18.400 --> 00:10:21.150 align:middle line:84% So at the family is [NON-ENGLISH SPEAKING].. 00:10:21.150 --> 00:10:22.380 align:middle line:90% You get "lah" for that one. 00:10:22.380 --> 00:10:25.057 align:middle line:90% 00:10:25.057 --> 00:10:25.640 align:middle line:90% Is this clear? 00:10:25.640 --> 00:10:27.200 align:middle line:90% That's the Finnish system. 00:10:27.200 --> 00:10:29.180 align:middle line:84% If any of you ever want to go learn Finnish, 00:10:29.180 --> 00:10:30.290 align:middle line:90% you're now all set up. 00:10:30.290 --> 00:10:31.415 align:middle line:90% Does anybody speak Finnish? 00:10:31.415 --> 00:10:33.035 align:middle line:84% Here I am pretending to speak Finnish. 00:10:33.035 --> 00:10:35.880 align:middle line:90% 00:10:35.880 --> 00:10:40.260 align:middle line:84% Yeah, so just to summarize that again, if a vowel is not round 00:10:40.260 --> 00:10:44.595 align:middle line:90% and not low, then it is front. 00:10:44.595 --> 00:10:49.030 align:middle line:90% 00:10:49.030 --> 00:10:52.720 align:middle line:84% So one way to think about what's going on in these suffix 00:10:52.720 --> 00:10:54.580 align:middle line:84% versions at the bottom is to say there's 00:10:54.580 --> 00:10:57.520 align:middle line:84% a rule saying give all the vowels in the word 00:10:57.520 --> 00:11:01.870 align:middle line:84% the same value for back, "o," except don't create vowels 00:11:01.870 --> 00:11:03.250 align:middle line:90% that don't exist in Finnish. 00:11:03.250 --> 00:11:10.300 align:middle line:84% So when you're doing "on the child," it would be nice if-- 00:11:10.300 --> 00:11:13.150 align:middle line:84% the fact is that you've got a back vowel and then 00:11:13.150 --> 00:11:15.880 align:middle line:90% a vowel and then a back vowel. 00:11:15.880 --> 00:11:19.530 align:middle line:84% But there's nothing you can do about the front vowel, eh. 00:11:19.530 --> 00:11:22.840 align:middle line:84% There isn't a corresponding back vowel in Finnish. 00:11:22.840 --> 00:11:25.220 align:middle line:90% So it's like the rules are-- 00:11:25.220 --> 00:11:31.500 align:middle line:84% first of all, remember Optimality Theory. 00:11:31.500 --> 00:11:34.350 align:middle line:84% So yeah, the generalization goes, 00:11:34.350 --> 00:11:36.360 align:middle line:84% give all the vowels in the word the same value 00:11:36.360 --> 00:11:39.240 align:middle line:84% for back, but don't create non-Finnish vowels. 00:11:39.240 --> 00:11:43.620 align:middle line:84% Don't create back vowels that are not round and not low. 00:11:43.620 --> 00:11:45.030 align:middle line:90% Don't create "ooh" or "eouh." 00:11:45.030 --> 00:11:47.557 align:middle line:84% There's that symbol you were wondering about. 00:11:47.557 --> 00:11:49.140 align:middle line:84% If we want to talk about this in terms 00:11:49.140 --> 00:11:50.515 align:middle line:84% of Optimality Theory, what we say 00:11:50.515 --> 00:11:54.810 align:middle line:84% is the most important thing is don't create back 00:11:54.810 --> 00:11:56.250 align:middle line:90% non-low unrounded vowels. 00:11:56.250 --> 00:11:57.900 align:middle line:90% Don't create "ooh" or "eouh." 00:11:57.900 --> 00:11:59.400 align:middle line:84% Finnish doesn't have those vowels. 00:11:59.400 --> 00:12:00.392 align:middle line:90% Don't do that. 00:12:00.392 --> 00:12:02.850 align:middle line:84% But as long as you-- so that's the most important priority. 00:12:02.850 --> 00:12:04.433 align:middle line:84% That's the highest ranking constraint. 00:12:04.433 --> 00:12:06.460 align:middle line:84% But as long as you don't do that, 00:12:06.460 --> 00:12:09.270 align:middle line:84% make sure all the vowels have the same value for back. 00:12:09.270 --> 00:12:11.387 align:middle line:84% They're all either back or front. 00:12:11.387 --> 00:12:13.720 align:middle line:84% That's one way to talk about what's going on in Finnish. 00:12:13.720 --> 00:12:18.070 align:middle line:90% 00:12:18.070 --> 00:12:21.430 align:middle line:84% Now there's a popular approach to phenomena like vowel harmony 00:12:21.430 --> 00:12:24.070 align:middle line:84% that I just wanted to introduce you to, partly 00:12:24.070 --> 00:12:26.560 align:middle line:84% because a version of it will float up again when we talk 00:12:26.560 --> 00:12:31.180 align:middle line:84% about syntax, I think, which involves saying 00:12:31.180 --> 00:12:33.620 align:middle line:84% it's as though-- and actually semantics as well-- 00:12:33.620 --> 00:12:40.840 align:middle line:84% it's as though we want to talk as though features of sounds 00:12:40.840 --> 00:12:42.853 align:middle line:84% can sometimes show up in multiple places. 00:12:42.853 --> 00:12:45.020 align:middle line:84% So here's the Finnish word for "on the table" again, 00:12:45.020 --> 00:12:46.900 align:middle line:90% [NON-ENGLISH SPEAKING]. 00:12:46.900 --> 00:12:52.410 align:middle line:84% And if you were listing all of the properties of all 00:12:52.410 --> 00:12:55.063 align:middle line:84% of those vowels-- that's what I've done here, 00:12:55.063 --> 00:12:56.730 align:middle line:84% and I just haven't listed the consonants 00:12:56.730 --> 00:12:58.800 align:middle line:84% just because there wasn't room and I'm not 00:12:58.800 --> 00:13:00.030 align:middle line:90% good at graphic design. 00:13:00.030 --> 00:13:06.930 align:middle line:84% But first of all, "ouh," is round and mid and not back. 00:13:06.930 --> 00:13:08.400 align:middle line:90% That is front. 00:13:08.400 --> 00:13:12.600 align:middle line:84% And the second vowel is round and high and not back. 00:13:12.600 --> 00:13:13.890 align:middle line:90% That is front. 00:13:13.890 --> 00:13:18.660 align:middle line:84% And "ae" is not round and is low, and is not back. 00:13:18.660 --> 00:13:19.840 align:middle line:90% That is, it's front. 00:13:19.840 --> 00:13:21.840 align:middle line:84% This is just a little list of all the properties 00:13:21.840 --> 00:13:24.570 align:middle line:90% of each of those vowels. 00:13:24.570 --> 00:13:27.960 align:middle line:84% And so you could specify all of these properties 00:13:27.960 --> 00:13:30.210 align:middle line:90% of these vowels this way. 00:13:30.210 --> 00:13:32.700 align:middle line:84% But there's a popular thought about what's 00:13:32.700 --> 00:13:37.530 align:middle line:84% going on here, which is to say it's as though Finnish wants 00:13:37.530 --> 00:13:42.570 align:middle line:84% the value of back to be kind of shared among all the vowels. 00:13:42.570 --> 00:13:44.640 align:middle line:84% That is, you're just going to state it once. 00:13:44.640 --> 00:13:48.510 align:middle line:84% This is a word in which all the vowels are front 00:13:48.510 --> 00:13:51.250 align:middle line:90% or all the vowels are back. 00:13:51.250 --> 00:13:55.650 align:middle line:84% So you say somewhere, Finnish doesn't allow you to state over 00:13:55.650 --> 00:13:57.942 align:middle line:84% and over again this vowel is front and this vowel front 00:13:57.942 --> 00:13:59.942 align:middle line:84% and this vowel is front and this vowel is front, 00:13:59.942 --> 00:14:02.040 align:middle line:84% you just have to say once, for the whole word, 00:14:02.040 --> 00:14:03.510 align:middle line:90% these vowels are front. 00:14:03.510 --> 00:14:05.160 align:middle line:84% And then there's this thing about, oh, 00:14:05.160 --> 00:14:06.868 align:middle line:84% and I don't like certain kinds of vowels, 00:14:06.868 --> 00:14:08.550 align:middle line:84% so those vowels get to be specified. 00:14:08.550 --> 00:14:13.260 align:middle line:84% But words just specify whether they're 00:14:13.260 --> 00:14:14.640 align:middle line:90% front words or back words. 00:14:14.640 --> 00:14:16.650 align:middle line:84% You don't specify it over and over again 00:14:16.650 --> 00:14:20.050 align:middle line:90% from one vowel to the next. 00:14:20.050 --> 00:14:21.993 align:middle line:84% This is one popular way of talking about how 00:14:21.993 --> 00:14:23.160 align:middle line:90% we want to do vowel harmony. 00:14:23.160 --> 00:14:26.350 align:middle line:90% 00:14:26.350 --> 00:14:28.890 align:middle line:84% It's as though there's only one back feature in the word, 00:14:28.890 --> 00:14:30.840 align:middle line:84% and all the vowels that can share it do. 00:14:30.840 --> 00:14:34.850 align:middle line:90% 00:14:34.850 --> 00:14:38.570 align:middle line:84% The name for this-- it's called autosegmental representation. 00:14:38.570 --> 00:14:41.000 align:middle line:84% This is the kind of moment where I 00:14:41.000 --> 00:14:45.770 align:middle line:84% think I'm swimming against the stream of your normal instincts 00:14:45.770 --> 00:14:46.310 align:middle line:90% for classes. 00:14:46.310 --> 00:14:50.180 align:middle line:84% This is a class in which there aren't quizzes or exams, 00:14:50.180 --> 00:14:52.940 align:middle line:84% so putting something like that up on the slide 00:14:52.940 --> 00:14:56.210 align:middle line:84% is triggering all your instincts to write that down and memorize 00:14:56.210 --> 00:14:58.970 align:middle line:84% it because you figure it'll come up on the test, 00:14:58.970 --> 00:15:00.810 align:middle line:90% but there is no test. 00:15:00.810 --> 00:15:03.148 align:middle line:84% So memorize that if you would like to, 00:15:03.148 --> 00:15:04.940 align:middle line:84% if you want to impress linguists at parties 00:15:04.940 --> 00:15:06.830 align:middle line:84% or if you want to make your TA happy, 00:15:06.830 --> 00:15:09.030 align:middle line:90% but there won't be a test. 00:15:09.030 --> 00:15:11.390 align:middle line:84% You will never be counted off on the paper 00:15:11.390 --> 00:15:13.970 align:middle line:84% for failing to use the phrase autosegmental representation. 00:15:13.970 --> 00:15:16.520 align:middle line:84% On the other hand, it is a way to impress people at parties. 00:15:16.520 --> 00:15:20.590 align:middle line:90% 00:15:20.590 --> 00:15:22.360 align:middle line:90% So that's one form of harmony. 00:15:22.360 --> 00:15:24.280 align:middle line:90% It's called vowel harmony. 00:15:24.280 --> 00:15:26.290 align:middle line:84% We've seen it now in Turkish and in Finnish. 00:15:26.290 --> 00:15:28.480 align:middle line:90% Those are not related languages. 00:15:28.480 --> 00:15:30.700 align:middle line:84% Vowel harmony is crosslinguistically not-- 00:15:30.700 --> 00:15:34.840 align:middle line:84% it's not hugely common, but it's not rare either. 00:15:34.840 --> 00:15:36.730 align:middle line:84% It's popular in the Turkic family. 00:15:36.730 --> 00:15:39.700 align:middle line:84% There's a very large family of languages 00:15:39.700 --> 00:15:41.620 align:middle line:84% that are related to each other and they all 00:15:41.620 --> 00:15:43.220 align:middle line:90% have various kinds of harmony. 00:15:43.220 --> 00:15:45.470 align:middle line:84% There's lots of interesting work on exactly what kinds 00:15:45.470 --> 00:15:47.360 align:middle line:84% of harmony you get in which languages. 00:15:47.360 --> 00:15:49.510 align:middle line:84% I want to show you another kind of harmony 00:15:49.510 --> 00:15:50.980 align:middle line:90% which is a lot rarer. 00:15:50.980 --> 00:15:53.260 align:middle line:84% This is from a language, the Navajo language, 00:15:53.260 --> 00:15:58.020 align:middle line:84% which is an Indigenous language of the American Southwest. 00:15:58.020 --> 00:16:00.690 align:middle line:84% Here are a bunch of verbs in Navajo 00:16:00.690 --> 00:16:04.740 align:middle line:84% in what we could think of as the past tense. 00:16:04.740 --> 00:16:08.670 align:middle line:84% And you can see that all of these verbs-- 00:16:08.670 --> 00:16:11.960 align:middle line:84% so what I've done is boldfaced a part of the verb 00:16:11.960 --> 00:16:14.460 align:middle line:84% that you can think of as the part of the verb that tells you 00:16:14.460 --> 00:16:15.720 align:middle line:90% who the subject is. 00:16:15.720 --> 00:16:20.070 align:middle line:84% So there's a morpheme "sé" that's showing up in the first 00:16:20.070 --> 00:16:24.660 align:middle line:84% three verbs, which tells you that the subject is "I." 00:16:24.660 --> 00:16:29.550 align:middle line:84% And there's a morpheme "síní" in the next trio of verbs that 00:16:29.550 --> 00:16:31.140 align:middle line:84% tells you that the subject is "you." 00:16:31.140 --> 00:16:35.040 align:middle line:84% And then there's a morpheme "soo" in the last three verbs. 00:16:35.040 --> 00:16:40.638 align:middle line:84% It tells you that the subject is "you two"-- that's you dual. 00:16:40.638 --> 00:16:42.180 align:middle line:84% To say that this is a single morpheme 00:16:42.180 --> 00:16:43.650 align:middle line:90% is a little bit misleading. 00:16:43.650 --> 00:16:45.390 align:middle line:84% The "s" at the beginning of this morpheme 00:16:45.390 --> 00:16:48.510 align:middle line:84% is probably partly a marker of what I'm translating here 00:16:48.510 --> 00:16:50.520 align:middle line:90% as past tense. 00:16:50.520 --> 00:16:54.660 align:middle line:84% Navajo verbs are extraordinarily complicated. 00:16:54.660 --> 00:16:58.050 align:middle line:84% So these particular verbs are in a class 00:16:58.050 --> 00:17:01.290 align:middle line:84% that takes this "s" version of the past tense, what 00:17:01.290 --> 00:17:02.400 align:middle line:90% are called s perfectives. 00:17:02.400 --> 00:17:03.720 align:middle line:90% There are also y perfectives. 00:17:03.720 --> 00:17:04.829 align:middle line:84% There are other kinds of perfectives. 00:17:04.829 --> 00:17:06.450 align:middle line:84% But anyway, this class of verbs takes 00:17:06.450 --> 00:17:09.410 align:middle line:84% these past tense morphemes, and this is what they look like. 00:17:09.410 --> 00:17:11.160 align:middle line:84% Are there any questions about any of this? 00:17:11.160 --> 00:17:11.996 align:middle line:90% Navajo has tone. 00:17:11.996 --> 00:17:13.829 align:middle line:84% That's what those markers are on the vowels. 00:17:13.829 --> 00:17:17.099 align:middle line:90% So the markers mark high tone. 00:17:17.099 --> 00:17:19.470 align:middle line:84% There's another marker that you're 00:17:19.470 --> 00:17:24.630 align:middle line:84% seeing underneath the vowel in the verb for "stand." 00:17:24.630 --> 00:17:26.670 align:middle line:90% That's a marker of nasalization. 00:17:26.670 --> 00:17:30.300 align:middle line:84% So "I stood" is something like [NON-ENGLISH SPEAKING].. 00:17:30.300 --> 00:17:34.620 align:middle line:84% So you have to nasalize the "eh," say that. 00:17:34.620 --> 00:17:35.490 align:middle line:90% Here's some Navajo. 00:17:35.490 --> 00:17:38.300 align:middle line:90% 00:17:38.300 --> 00:17:40.520 align:middle line:90% Here's some more Navajo. 00:17:40.520 --> 00:17:43.730 align:middle line:84% "I boiled it" is [NON-ENGLISH SPEAKING].. 00:17:43.730 --> 00:17:46.280 align:middle line:84% "You boiled it" is [NON-ENGLISH SPEAKING],, 00:17:46.280 --> 00:17:51.080 align:middle line:84% and "You two boiled it" is [NON-ENGLISH SPEAKING].. 00:17:51.080 --> 00:17:54.830 align:middle line:84% So you can see, in the three verbs that I had 00:17:54.830 --> 00:17:57.500 align:middle line:84% on the previous slide and that are still up here, 00:17:57.500 --> 00:18:01.270 align:middle line:84% there are these three morphemes, "sé," "síní," and "soo." 00:18:01.270 --> 00:18:04.200 align:middle line:84% But for "boil," the morphemes are [NON-ENGLISH SPEAKING] 00:18:04.200 --> 00:18:07.200 align:middle line:90% and [NON-ENGLISH SPEAKING]. 00:18:07.200 --> 00:18:12.130 align:middle line:90% This is called sibilant harmony. 00:18:12.130 --> 00:18:16.180 align:middle line:84% What's going on here-- and here's another example of it-- 00:18:16.180 --> 00:18:20.740 align:middle line:84% you get "sé," "síní," and "soo" for "played" and "stood" 00:18:20.740 --> 00:18:23.560 align:middle line:84% and "crawled around," but [NON-ENGLISH SPEAKING] 00:18:23.560 --> 00:18:27.340 align:middle line:84% and [NON-ENGLISH SPEAKING] for "boiled" and "worked." 00:18:27.340 --> 00:18:30.550 align:middle line:84% And the relevant difference between "boil" and "work" 00:18:30.550 --> 00:18:32.890 align:middle line:84% on the one hand and "play," "crawl around," 00:18:32.890 --> 00:18:35.770 align:middle line:84% and "stand" on the other is that "boil" and "work" 00:18:35.770 --> 00:18:40.420 align:middle line:84% have, in the verb, a sibilant, which is postalveolar. 00:18:40.420 --> 00:18:42.790 align:middle line:84% So "boil"-- the stem you can think 00:18:42.790 --> 00:18:45.370 align:middle line:84% of as [NON-ENGLISH SPEAKING],, and "work"-- 00:18:45.370 --> 00:18:47.710 align:middle line:84% its' [NON-ENGLISH SPEAKING] is the verb stem, 00:18:47.710 --> 00:18:49.600 align:middle line:90% [NON-ENGLISH SPEAKING]. 00:18:49.600 --> 00:18:54.160 align:middle line:84% So "boil" has a "zh" in it and "work" has a "sh" in it, 00:18:54.160 --> 00:18:57.115 align:middle line:84% whereas "play" and "crawl around" and "stand" either 00:18:57.115 --> 00:18:59.740 align:middle line:84% don't have sibilance in them at all or, in the case of "stand," 00:18:59.740 --> 00:19:02.170 align:middle line:84% it has a sibilant, but it's not a postalveolar sibilant, 00:19:02.170 --> 00:19:05.440 align:middle line:90% It's an alveolar sibilant, "z." 00:19:05.440 --> 00:19:07.475 align:middle line:84% This is called sibilant harmony, or I 00:19:07.475 --> 00:19:09.100 align:middle line:84% think we've also called these strident. 00:19:09.100 --> 00:19:11.120 align:middle line:84% I'm sorry, that's the term you're familiar with. 00:19:11.120 --> 00:19:14.890 align:middle line:84% So this phenomenon is called sibilant harmony-- 00:19:14.890 --> 00:19:17.860 align:middle line:84% so sibilant is another name for strident, sorry. 00:19:17.860 --> 00:19:19.640 align:middle line:84% So let me just say all that again, 00:19:19.640 --> 00:19:23.260 align:middle line:84% there are these postalveolar stridents, "zh" and "sh" 00:19:23.260 --> 00:19:26.260 align:middle line:84% in "boil" and "work," and those are triggering the postalveolar 00:19:26.260 --> 00:19:28.660 align:middle line:84% versions of these morphemes, the [NON-ENGLISH SPEAKING] 00:19:28.660 --> 00:19:31.030 align:middle line:84% and [NON-ENGLISH SPEAKING] instead of "sé," "síní," 00:19:31.030 --> 00:19:33.690 align:middle line:90% and "soo." 00:19:33.690 --> 00:19:36.787 align:middle line:84% I started this by saying vowel harmony is pretty common. 00:19:36.787 --> 00:19:38.370 align:middle line:84% There's a lot of work on vowel harmony 00:19:38.370 --> 00:19:40.495 align:middle line:84% because you find it in lots of unrelated languages. 00:19:40.495 --> 00:19:42.840 align:middle line:90% Sibilant harmony is quite rare. 00:19:42.840 --> 00:19:43.520 align:middle line:90% Navajo has it. 00:19:43.520 --> 00:19:45.150 align:middle line:90% A few other languages have it. 00:19:45.150 --> 00:19:51.330 align:middle line:84% But it's similar in the sense that it looks as though, 00:19:51.330 --> 00:19:53.940 align:middle line:90% if you have a sibling-- 00:19:53.940 --> 00:19:56.460 align:middle line:84% morphemes in these prefixes whose 00:19:56.460 --> 00:20:02.580 align:middle line:84% form is being conditioned by corresponding sounds 00:20:02.580 --> 00:20:03.420 align:middle line:90% in the verb stem. 00:20:03.420 --> 00:20:05.910 align:middle line:84% So for vowel harmony, it's the vowels. 00:20:05.910 --> 00:20:08.100 align:middle line:84% For sibilant harmony, it's these stridents. 00:20:08.100 --> 00:20:12.240 align:middle line:84% So if there's a strident or sibilant, an "s," "z," "sh," 00:20:12.240 --> 00:20:14.610 align:middle line:84% or "zh"-- those are the Navajo sibilants-- 00:20:14.610 --> 00:20:17.130 align:middle line:84% in the verb stem, then any of the sibilance and the subject 00:20:17.130 --> 00:20:19.198 align:middle line:90% prefix have to match it. 00:20:19.198 --> 00:20:20.490 align:middle line:90% They're getting the same thing. 00:20:20.490 --> 00:20:25.580 align:middle line:90% 00:20:25.580 --> 00:20:27.170 align:middle line:84% Vowel harmony and maybe this idea 00:20:27.170 --> 00:20:29.767 align:middle line:84% of sharing a feature across sounds-- 00:20:29.767 --> 00:20:31.100 align:middle line:90% it's not just for vowel harmony. 00:20:31.100 --> 00:20:32.830 align:middle line:84% Maybe it's for assimilation generally. 00:20:32.830 --> 00:20:34.580 align:middle line:84% There might be some other kinds of harmony 00:20:34.580 --> 00:20:37.700 align:middle line:90% we'd want to apply it to. 00:20:37.700 --> 00:20:40.160 align:middle line:84% Wanted to show you another bit of Navajo 00:20:40.160 --> 00:20:43.710 align:middle line:90% before we leave Navajo behind. 00:20:43.710 --> 00:20:46.940 align:middle line:84% Here's how you say you are playing in Navajo. 00:20:46.940 --> 00:20:48.547 align:middle line:90% So we're switching tenses now. 00:20:48.547 --> 00:20:50.630 align:middle line:84% We're going to get away from sibilant and harmony. 00:20:50.630 --> 00:20:54.800 align:middle line:84% This is the present tense, the imperfective aspect people 00:20:54.800 --> 00:20:56.330 align:middle line:84% call it when they work on Navajo. 00:20:56.330 --> 00:20:57.378 align:middle line:90% It's not exactly tense. 00:20:57.378 --> 00:20:59.420 align:middle line:84% So this is how you say you are playing in Navajo. 00:20:59.420 --> 00:21:01.507 align:middle line:90% It's [NON-ENGLISH SPEAKING]. 00:21:01.507 --> 00:21:04.720 align:middle line:90% 00:21:04.720 --> 00:21:08.020 align:middle line:84% This verb has three morphemes in it. 00:21:08.020 --> 00:21:10.960 align:middle line:84% You can think of the "né" at the end as the part that really 00:21:10.960 --> 00:21:12.820 align:middle line:90% means "play." 00:21:12.820 --> 00:21:15.620 align:middle line:84% There's a morpheme at the beginning, "na-," 00:21:15.620 --> 00:21:18.580 align:middle line:90% which is a prefix. 00:21:18.580 --> 00:21:21.370 align:middle line:84% We don't have to worry for this about what it means. 00:21:21.370 --> 00:21:23.560 align:middle line:84% It's a marker of a certain class of verbs 00:21:23.560 --> 00:21:24.970 align:middle line:90% what are called atelic verbs. 00:21:24.970 --> 00:21:28.900 align:middle line:84% These are verbs in which you can continue doing something 00:21:28.900 --> 00:21:32.210 align:middle line:84% and there is no natural endpoint to what you do. 00:21:32.210 --> 00:21:37.000 align:middle line:84% So if you compare something like walking around 00:21:37.000 --> 00:21:39.460 align:middle line:90% to walking to work-- 00:21:39.460 --> 00:21:44.300 align:middle line:84% if you walk to work, then when you get to work you stop. 00:21:44.300 --> 00:21:47.770 align:middle line:84% And if you interrupt me while I'm 00:21:47.770 --> 00:21:49.270 align:middle line:84% in the middle of walking to work, 00:21:49.270 --> 00:21:51.730 align:middle line:90% then I did not walk to work. 00:21:51.730 --> 00:21:56.770 align:middle line:84% That's a telic predicate, as opposed to walking around. 00:21:56.770 --> 00:21:59.500 align:middle line:84% If I'm walking around for an hour and then you stop me, 00:21:59.500 --> 00:22:01.780 align:middle line:84% well, I still walked around for a while. 00:22:01.780 --> 00:22:02.860 align:middle line:90% So that's atelic. 00:22:02.860 --> 00:22:06.160 align:middle line:84% It doesn't have a natural endpoint, something 00:22:06.160 --> 00:22:07.370 align:middle line:90% you're trying to achieve. 00:22:07.370 --> 00:22:12.527 align:middle line:84% So "na-" is a marker of atelic verbs like play or walk around. 00:22:12.527 --> 00:22:14.360 align:middle line:84% Feel free to ignore all that if you want to. 00:22:14.360 --> 00:22:16.160 align:middle line:84% So there's a "na-" at the beginning of this verb. 00:22:16.160 --> 00:22:18.220 align:middle line:84% There's a "né" at the end, which is the part that really means 00:22:18.220 --> 00:22:18.730 align:middle line:90% play. 00:22:18.730 --> 00:22:20.280 align:middle line:84% And then in between them, there's 00:22:20.280 --> 00:22:22.810 align:middle line:84% an agreement morpheme, something that's telling you 00:22:22.810 --> 00:22:24.460 align:middle line:90% what the subject is. 00:22:24.460 --> 00:22:29.080 align:middle line:84% And that morpheme is "ni" for you, and it's "sh" for I, 00:22:29.080 --> 00:22:33.280 align:middle line:84% and it's "oh" for you two, and it's "ii'" for we two. 00:22:33.280 --> 00:22:38.366 align:middle line:84% So you get [NON-ENGLISH SPEAKING] 00:22:38.366 --> 00:22:40.665 align:middle line:90% for those forms of the verb. 00:22:40.665 --> 00:22:42.040 align:middle line:84% So far, here's a bunch of Navajo. 00:22:42.040 --> 00:22:45.160 align:middle line:90% 00:22:45.160 --> 00:22:48.940 align:middle line:84% Oh, one thing I have to ask you to believe me about because it 00:22:48.940 --> 00:22:51.840 align:middle line:84% would be hard to demonstrate is that the ee in we 00:22:51.840 --> 00:22:55.030 align:middle line:84% two are playing is underlying-- so that apostrophe there, 00:22:55.030 --> 00:22:56.350 align:middle line:90% this is Navajo orthography. 00:22:56.350 --> 00:22:59.620 align:middle line:84% It represents a glottal stop, but that glottal stop 00:22:59.620 --> 00:23:00.910 align:middle line:90% is originally a "d." 00:23:00.910 --> 00:23:03.160 align:middle line:84% I'll just ask you to trust me about that 00:23:03.160 --> 00:23:05.900 align:middle line:90% for the next little while. 00:23:05.900 --> 00:23:12.330 align:middle line:84% Now I started with play, which is a comparatively simple verb. 00:23:12.330 --> 00:23:14.960 align:middle line:90% The verb really is "né." 00:23:14.960 --> 00:23:17.480 align:middle line:84% But there are some other verbs that we can talk about. 00:23:17.480 --> 00:23:21.560 align:middle line:84% You've seen versions of them in the preceding slides-- 00:23:21.560 --> 00:23:26.600 align:middle line:84% work and investigate-- that I now want to turn to. 00:23:26.600 --> 00:23:29.330 align:middle line:84% Work and investigate-- we can think of them 00:23:29.330 --> 00:23:31.340 align:middle line:84% as starting with clusters of consonants. 00:23:31.340 --> 00:23:33.380 align:middle line:90% So you are working. 00:23:33.380 --> 00:23:36.590 align:middle line:84% You've got the "na" and the "ni" that we have up above, 00:23:36.590 --> 00:23:41.030 align:middle line:84% and then the verb stem is [NON-ENGLISH SPEAKING].. 00:23:41.030 --> 00:23:43.640 align:middle line:84% People who work on Navajo really think that that "l" 00:23:43.640 --> 00:23:45.630 align:middle line:90% is a separate morpheme. 00:23:45.630 --> 00:23:48.650 align:middle line:90% They call it the classifier. 00:23:48.650 --> 00:23:50.690 align:middle line:84% Really, for our purposes, all that amounts to 00:23:50.690 --> 00:23:53.360 align:middle line:84% is that there are a bunch of verbs that start with "l," 00:23:53.360 --> 00:23:55.880 align:middle line:84% and then the "l" is followed by another consonant. 00:23:55.880 --> 00:23:58.920 align:middle line:84% We don't have to worry too much about what that is. 00:23:58.920 --> 00:24:01.970 align:middle line:84% So for our purposes, "working" starts with two consonants, 00:24:01.970 --> 00:24:02.990 align:middle line:90% [NON-ENGLISH SPEAKING]. 00:24:02.990 --> 00:24:06.300 align:middle line:84% And "investigating" also starts with two consonants. 00:24:06.300 --> 00:24:09.460 align:middle line:84% The first one is a voiceless [NON-ENGLISH SPEAKING] so this 00:24:09.460 --> 00:24:11.750 align:middle line:84% [NON-ENGLISH SPEAKING] is "You are investigating it." 00:24:11.750 --> 00:24:15.430 align:middle line:90% 00:24:15.430 --> 00:24:16.890 align:middle line:90% Why am I telling you all this? 00:24:16.890 --> 00:24:20.140 align:middle line:90% This is where the action is. 00:24:20.140 --> 00:24:26.140 align:middle line:84% "Ni-" is the prefix for you singular, 00:24:26.140 --> 00:24:28.390 align:middle line:90% and it ends with a vowel. 00:24:28.390 --> 00:24:31.450 align:middle line:84% But if we switch to the other prefixes, which 00:24:31.450 --> 00:24:35.630 align:middle line:84% all end in consonants, then some interesting things happen. 00:24:35.630 --> 00:24:37.600 align:middle line:90% So I am playing-- 00:24:37.600 --> 00:24:40.900 align:middle line:84% we saw that the subject agreement prefix for "I" 00:24:40.900 --> 00:24:42.850 align:middle line:90% is a "sh." 00:24:42.850 --> 00:24:44.560 align:middle line:84% So you get [NON-ENGLISH SPEAKING].. 00:24:44.560 --> 00:24:46.460 align:middle line:84% The subject agreement prefix "sh," 00:24:46.460 --> 00:24:48.460 align:middle line:84% if you put it before it [NON-ENGLISH SPEAKING],, 00:24:48.460 --> 00:24:50.918 align:middle line:84% you might have expected that to be [NON-ENGLISH SPEAKING].. 00:24:50.918 --> 00:24:51.640 align:middle line:90% But it isn't. 00:24:51.640 --> 00:24:52.810 align:middle line:90% It's [NON-ENGLISH SPEAKING]. 00:24:52.810 --> 00:24:54.940 align:middle line:90% So the "l" vanishes. 00:24:54.940 --> 00:24:59.945 align:middle line:84% Similarly for investigate, we've decided that "investigate" 00:24:59.945 --> 00:25:01.300 align:middle line:90% is [NON-ENGLISH SPEAKING]. 00:25:01.300 --> 00:25:03.610 align:middle line:84% So it starts with two consonants, a voiceless "l" 00:25:03.610 --> 00:25:04.340 align:middle line:90% and then a "k." 00:25:04.340 --> 00:25:06.340 align:middle line:84% If you also have a [NON-ENGLISH SPEAKING] before 00:25:06.340 --> 00:25:08.440 align:middle line:84% all that, well, you get rid of the voiceless "l." 00:25:08.440 --> 00:25:09.940 align:middle line:84% So you get [NON-ENGLISH SPEAKING] is 00:25:09.940 --> 00:25:10.960 align:middle line:90% "I'm investigating it." 00:25:10.960 --> 00:25:16.980 align:middle line:90% 00:25:16.980 --> 00:25:19.950 align:middle line:84% The reason I'm doing these particular prefixes 00:25:19.950 --> 00:25:22.720 align:middle line:84% is that you can see different things with the different ones. 00:25:22.720 --> 00:25:28.620 align:middle line:84% So if you add "oh," what happens is that the "h" goes away, 00:25:28.620 --> 00:25:31.650 align:middle line:84% but the distinction between the voiced "l" in "work" 00:25:31.650 --> 00:25:34.750 align:middle line:84% and the voiceless "l" in "investigate" also goes away. 00:25:34.750 --> 00:25:37.500 align:middle line:84% So "work"-- we decided, looking at you, 00:25:37.500 --> 00:25:40.440 align:middle line:84% starts with a voiced "l" It's [NON-ENGLISH SPEAKING].. 00:25:40.440 --> 00:25:42.990 align:middle line:84% But "You two are working" is now [NON-ENGLISH SPEAKING].. 00:25:42.990 --> 00:25:46.110 align:middle line:84% So the "h" is gone, but that voiced 00:25:46.110 --> 00:25:47.573 align:middle line:90% "l" has become voiceless. 00:25:47.573 --> 00:25:48.240 align:middle line:90% It's now barred. 00:25:48.240 --> 00:25:50.517 align:middle line:90% 00:25:50.517 --> 00:25:52.350 align:middle line:84% And we're getting a voiceless "l" down there 00:25:52.350 --> 00:25:53.933 align:middle line:84% in "You two are investigating it" too. 00:25:53.933 --> 00:25:56.390 align:middle line:90% It's now [NON-ENGLISH SPEAKING]. 00:25:56.390 --> 00:26:02.400 align:middle line:84% And the thing that I said is underlyingly "eed," well, 00:26:02.400 --> 00:26:04.557 align:middle line:84% the "d" is gone, but the distinction 00:26:04.557 --> 00:26:06.390 align:middle line:84% between the voiced "l" and the voiceless "l" 00:26:06.390 --> 00:26:08.290 align:middle line:90% has gone in the other direction. 00:26:08.290 --> 00:26:09.630 align:middle line:90% So now they're both voiced. 00:26:09.630 --> 00:26:12.240 align:middle line:84% So "We two are working" is now [NON-ENGLISH SPEAKING] and "We 00:26:12.240 --> 00:26:14.580 align:middle line:84% two are investigating" is now [NON-ENGLISH SPEAKING].. 00:26:14.580 --> 00:26:16.622 align:middle line:84% Do you remember when I said that Navajo verbs are 00:26:16.622 --> 00:26:18.090 align:middle line:90% extraordinarily complicated? 00:26:18.090 --> 00:26:20.230 align:middle line:84% Navajo verbs are extraordinarily complicated. 00:26:20.230 --> 00:26:22.440 align:middle line:84% There are all these morphemes, but they 00:26:22.440 --> 00:26:25.530 align:middle line:90% don't coexist peacefully. 00:26:25.530 --> 00:26:28.350 align:middle line:84% When we were talking about Native languages in which you 00:26:28.350 --> 00:26:29.640 align:middle line:84% have a whole bunch of morphemes, and they're 00:26:29.640 --> 00:26:30.990 align:middle line:84% easy to separate from each other, 00:26:30.990 --> 00:26:32.790 align:middle line:84% and there's nothing very spectacular that happens, 00:26:32.790 --> 00:26:34.290 align:middle line:84% you just attach them to each other-- 00:26:34.290 --> 00:26:36.930 align:middle line:84% this is the natural enemy of that kind of language. 00:26:36.930 --> 00:26:37.950 align:middle line:90% It's the opposite. 00:26:37.950 --> 00:26:39.727 align:middle line:84% Here, we have a bunch of morphemes, 00:26:39.727 --> 00:26:41.310 align:middle line:84% and you put them together, and there's 00:26:41.310 --> 00:26:42.660 align:middle line:90% like open warfare between them. 00:26:42.660 --> 00:26:44.010 align:middle line:84% There's all kinds of strange stuff 00:26:44.010 --> 00:26:45.927 align:middle line:84% that happens as they interact with each other. 00:26:45.927 --> 00:26:48.720 align:middle line:90% 00:26:48.720 --> 00:26:51.800 align:middle line:84% So let me summarize what we've just seen. 00:26:51.800 --> 00:26:54.860 align:middle line:84% If you have the "sh" followed by two consonants, 00:26:54.860 --> 00:26:57.590 align:middle line:84% like if you're going to say "I am investigating it," so "sh" 00:26:57.590 --> 00:27:00.148 align:middle line:84% is the "I," and "investigating," we decided, 00:27:00.148 --> 00:27:01.940 align:middle line:84% starts with [NON-ENGLISH SPEAKING] and then 00:27:01.940 --> 00:27:06.930 align:middle line:84% "k," you get rid of the first consonant. 00:27:06.930 --> 00:27:08.930 align:middle line:84% So you get [NON-ENGLISH SPEAKING] not 00:27:08.930 --> 00:27:11.700 align:middle line:90% [NON-ENGLISH SPEAKING]. 00:27:11.700 --> 00:27:16.620 align:middle line:84% If you have an "h" followed by two consonants, 00:27:16.620 --> 00:27:21.180 align:middle line:84% like in "We two are working," so where you've got an "oh" 00:27:21.180 --> 00:27:23.850 align:middle line:84% followed by [NON-ENGLISH SPEAKING],, well, 00:27:23.850 --> 00:27:27.600 align:middle line:84% the "h" goes away, but the "l" devoices. 00:27:27.600 --> 00:27:30.270 align:middle line:84% So instead of [NON-ENGLISH SPEAKING],, 00:27:30.270 --> 00:27:33.310 align:middle line:90% we get [NON-ENGLISH SPEAKING]. 00:27:33.310 --> 00:27:38.130 align:middle line:84% And if you have a "d" before two consonants, well, 00:27:38.130 --> 00:27:42.900 align:middle line:84% the "d" goes away, but the "l" that's in the first consonant 00:27:42.900 --> 00:27:44.250 align:middle line:90% will become voiced. 00:27:44.250 --> 00:27:47.250 align:middle line:84% So it changes from [NON-ENGLISH SPEAKING]---- 00:27:47.250 --> 00:27:51.402 align:middle line:84% so "You two are investigating it" to [NON-ENGLISH SPEAKING].. 00:27:51.402 --> 00:27:54.120 align:middle line:90% 00:27:54.120 --> 00:28:00.370 align:middle line:84% Do these rules remind you of anything other than the Navajo 00:28:00.370 --> 00:28:03.370 align:middle line:90% that we've just gone through? 00:28:03.370 --> 00:28:07.950 align:middle line:84% So we have one rule that says if you have a "sh" followed by two 00:28:07.950 --> 00:28:10.330 align:middle line:84% consonants, get rid of the first consonant, 00:28:10.330 --> 00:28:13.350 align:middle line:84% and another rule that says, if you have an "h" followed 00:28:13.350 --> 00:28:15.450 align:middle line:84% by two consonants, get rid of the "h" 00:28:15.450 --> 00:28:18.680 align:middle line:84% but devoice the first consonant, and then another rule 00:28:18.680 --> 00:28:21.680 align:middle line:84% that says, if you have a "d" followed by two consonants, 00:28:21.680 --> 00:28:26.030 align:middle line:84% get rid of the "d," oh, but voice the first consonant. 00:28:26.030 --> 00:28:26.982 align:middle line:90% Yeah? 00:28:26.982 --> 00:28:28.940 align:middle line:84% AUDIENCE: I might be barking up the wrong tree, 00:28:28.940 --> 00:28:32.813 align:middle line:84% but was there a language that disallowed [INAUDIBLE]?? 00:28:32.813 --> 00:28:34.230 align:middle line:84% NORVIN RICHARDS: There was indeed. 00:28:34.230 --> 00:28:35.772 align:middle line:84% You're not barking up the wrong tree. 00:28:35.772 --> 00:28:37.410 align:middle line:84% You're barking up the correct tree. 00:28:37.410 --> 00:28:40.050 align:middle line:84% It's not something I've ever had to say before, but it's true. 00:28:40.050 --> 00:28:42.340 align:middle line:84% You are barking up the Yawelmani tree. 00:28:42.340 --> 00:28:44.910 align:middle line:84% So we were talking about Yawelmani before, 00:28:44.910 --> 00:28:47.490 align:middle line:84% and for Yawelmani, similarly, we had a whole bunch 00:28:47.490 --> 00:28:49.890 align:middle line:84% of phonological phenomena, and we 00:28:49.890 --> 00:28:52.230 align:middle line:84% said we could write rules to handle all of these facts 00:28:52.230 --> 00:28:54.540 align:middle line:90% and maybe we should. 00:28:54.540 --> 00:28:55.915 align:middle line:84% In Yawelmani, it was like, if you 00:28:55.915 --> 00:28:57.832 align:middle line:84% have three consonants in a row and one of them 00:28:57.832 --> 00:28:59.250 align:middle line:90% is an "h," get rid of the "h." 00:28:59.250 --> 00:29:00.720 align:middle line:84% If you have three consonants in a row and none of them 00:29:00.720 --> 00:29:02.400 align:middle line:90% is an "h," you insert a vowel. 00:29:02.400 --> 00:29:04.110 align:middle line:84% There are various things that Yawelmani 00:29:04.110 --> 00:29:06.390 align:middle line:84% does to avoid having three consonants in a row. 00:29:06.390 --> 00:29:09.900 align:middle line:84% Navajo is doing the same thing, you're right-- 00:29:09.900 --> 00:29:12.120 align:middle line:84% the same thing in the sense that Navajo and Yawelmani 00:29:12.120 --> 00:29:15.060 align:middle line:84% both dislike sequences of three consonants in a row. 00:29:15.060 --> 00:29:18.780 align:middle line:84% None of these rules would handle anything in Yawelmani. 00:29:18.780 --> 00:29:20.820 align:middle line:84% So this is not what Yawelmani does to deal 00:29:20.820 --> 00:29:22.448 align:middle line:90% with three consonants in a row. 00:29:22.448 --> 00:29:23.865 align:middle line:84% But if we're willing to think of-- 00:29:23.865 --> 00:29:26.560 align:middle line:90% 00:29:26.560 --> 00:29:28.560 align:middle line:84% we're willing to think of phonology the way I've 00:29:28.560 --> 00:29:31.080 align:middle line:84% been advertising, which is languages 00:29:31.080 --> 00:29:33.030 align:middle line:84% have certain problems, and then they 00:29:33.030 --> 00:29:34.950 align:middle line:84% have means to deal with the problems, 00:29:34.950 --> 00:29:36.660 align:middle line:84% then we can see Yawelmani and Navajo 00:29:36.660 --> 00:29:38.020 align:middle line:90% as having something in common. 00:29:38.020 --> 00:29:40.270 align:middle line:84% They have a kind of problem that they both care about. 00:29:40.270 --> 00:29:42.240 align:middle line:84% They don't like sequences of three consonants. 00:29:42.240 --> 00:29:44.490 align:middle line:84% They deal with it in different ways, 00:29:44.490 --> 00:29:47.680 align:middle line:84% but we can see them as having that property in common. 00:29:47.680 --> 00:29:48.180 align:middle line:90% Yes? 00:29:48.180 --> 00:29:49.440 align:middle line:84% AUDIENCE: Is this a property that's 00:29:49.440 --> 00:29:50.815 align:middle line:84% been observed in other Indigenous 00:29:50.815 --> 00:29:52.895 align:middle line:84% languages of the Americas, or [INAUDIBLE]?? 00:29:52.895 --> 00:29:54.270 align:middle line:84% NORVIN RICHARDS: I think it's not 00:29:54.270 --> 00:29:56.580 align:middle line:84% specific to Indigenous languages of the Americas. 00:29:56.580 --> 00:29:57.990 align:middle line:84% I think it's cross-linguistically 00:29:57.990 --> 00:30:00.000 align:middle line:84% not uncommon for languages to dislike 00:30:00.000 --> 00:30:03.360 align:middle line:90% sequences of three consonants. 00:30:03.360 --> 00:30:07.860 align:middle line:84% And we talked a little bit earlier on about the fact 00:30:07.860 --> 00:30:12.510 align:middle line:84% that some of your best cues for the value of a consonant 00:30:12.510 --> 00:30:15.840 align:middle line:84% is its effects on the nearby vowel. 00:30:15.840 --> 00:30:20.070 align:middle line:84% So I think we played around with [INAUDIBLE],, 00:30:20.070 --> 00:30:21.990 align:middle line:84% showed you some spectrograms from me saying 00:30:21.990 --> 00:30:23.700 align:middle line:90% "bab" and "dad" and "gag." 00:30:23.700 --> 00:30:26.325 align:middle line:84% I showed you that during the stop itself, 00:30:26.325 --> 00:30:27.450 align:middle line:90% you can't hear a whole lot. 00:30:27.450 --> 00:30:30.690 align:middle line:84% It's what the stop does to the vowel that's nearby. 00:30:30.690 --> 00:30:32.340 align:middle line:84% If you have three consonants in a row, 00:30:32.340 --> 00:30:35.310 align:middle line:84% then at least the one in the middle is not next to a vowel, 00:30:35.310 --> 00:30:37.680 align:middle line:84% and so that could be a reason why 00:30:37.680 --> 00:30:39.720 align:middle line:84% it's bad to have three consonants in a row. 00:30:39.720 --> 00:30:42.867 align:middle line:84% It's perfectly possible there are 00:30:42.867 --> 00:30:44.700 align:middle line:84% languages out there that are fine with three 00:30:44.700 --> 00:30:45.540 align:middle line:90% consonants in a row. 00:30:45.540 --> 00:30:47.170 align:middle line:84% English is one of them for example. 00:30:47.170 --> 00:30:50.160 align:middle line:84% So we have words like "strengths"-- 00:30:50.160 --> 00:30:52.320 align:middle line:90% long strings of consonants. 00:30:52.320 --> 00:30:53.820 align:middle line:84% Plenty of other languages like that. 00:30:53.820 --> 00:30:56.590 align:middle line:84% But yeah, this is a kind of thing that languages do. 00:30:56.590 --> 00:30:59.820 align:middle line:84% And we're seeing it in Yawelmani and Navajo. 00:30:59.820 --> 00:31:02.760 align:middle line:84% So this is one virtue of the approach 00:31:02.760 --> 00:31:05.010 align:middle line:84% to phonology that I was advertising before, 00:31:05.010 --> 00:31:08.795 align:middle line:84% where we say, yeah, languages have things they dislike, 00:31:08.795 --> 00:31:10.170 align:middle line:84% and then, on the other hand, they 00:31:10.170 --> 00:31:12.810 align:middle line:84% have techniques for dealing with what they dislike. 00:31:12.810 --> 00:31:14.550 align:middle line:84% Yawelmani and Navajo have a property 00:31:14.550 --> 00:31:16.440 align:middle line:84% in common, namely the property of disliking 00:31:16.440 --> 00:31:17.610 align:middle line:90% three consonants in a row. 00:31:17.610 --> 00:31:19.830 align:middle line:84% They deal with that problem in different ways, 00:31:19.830 --> 00:31:22.420 align:middle line:84% but they have that property in common, at least. 00:31:22.420 --> 00:31:24.237 align:middle line:84% And so there's some kind of insight 00:31:24.237 --> 00:31:26.070 align:middle line:84% that we're gaining by being willing to think 00:31:26.070 --> 00:31:27.537 align:middle line:90% about phonology that way. 00:31:27.537 --> 00:31:29.370 align:middle line:84% If we didn't think about phonology that way, 00:31:29.370 --> 00:31:33.810 align:middle line:84% if we just made these rules, Navajo and Yawelmani 00:31:33.810 --> 00:31:36.900 align:middle line:84% would seem not to have any properties in common at all. 00:31:36.900 --> 00:31:38.610 align:middle line:84% We'd obscure that fact about them. 00:31:38.610 --> 00:31:39.330 align:middle line:90% That make sense? 00:31:39.330 --> 00:31:46.290 align:middle line:90% 00:31:46.290 --> 00:31:48.700 align:middle line:90% So what have we done today-- 00:31:48.700 --> 00:31:51.180 align:middle line:84% we've looked at a variety of phenomena. 00:31:51.180 --> 00:31:52.920 align:middle line:84% I tried to introduce you to the idea 00:31:52.920 --> 00:31:54.900 align:middle line:84% of autosegmental representation, this idea 00:31:54.900 --> 00:31:57.690 align:middle line:84% that a single feature can get spread 00:31:57.690 --> 00:31:59.880 align:middle line:84% across different places in a word, 00:31:59.880 --> 00:32:04.500 align:middle line:84% and we've just gone back to talking about this idea of what 00:32:04.500 --> 00:32:07.560 align:middle line:84% are called ranked constraints, this idea that languages might 00:32:07.560 --> 00:32:13.088 align:middle line:84% have a number of goals in life, and the way 00:32:13.088 --> 00:32:14.880 align:middle line:84% to talk about differences between languages 00:32:14.880 --> 00:32:17.340 align:middle line:84% might just be in terms of how they prioritize 00:32:17.340 --> 00:32:18.340 align:middle line:90% their different goals. 00:32:18.340 --> 00:32:19.840 align:middle line:84% So that was an important part of how 00:32:19.840 --> 00:32:22.260 align:middle line:84% we talked about Finnish, where the rules seem 00:32:22.260 --> 00:32:25.170 align:middle line:84% to be do vowel harmony, oh, but don't 00:32:25.170 --> 00:32:26.627 align:middle line:90% create non-Finnish vowels. 00:32:26.627 --> 00:32:28.710 align:middle line:84% So the most important thing is don't create vowels 00:32:28.710 --> 00:32:31.560 align:middle line:84% that Finnish doesn't have, but modulo that, make 00:32:31.560 --> 00:32:36.360 align:middle line:84% all your vowels the same for whether they're front or back. 00:32:36.360 --> 00:32:38.328 align:middle line:84% And similarly, we just convinced ourselves 00:32:38.328 --> 00:32:40.620 align:middle line:84% that Navajo and Yawelmani, although they look different 00:32:40.620 --> 00:32:43.140 align:middle line:84% in lots of other ways, might have a high-ranking constraint 00:32:43.140 --> 00:32:47.340 align:middle line:84% in common, namely, don't have three consonants in a row. 00:32:47.340 --> 00:32:51.180 align:middle line:84% So what I want to do now is show you 00:32:51.180 --> 00:32:54.930 align:middle line:84% another domain in which people have fruitfully 00:32:54.930 --> 00:32:57.960 align:middle line:84% made use of this idea of ranked constraints-- that is, 00:32:57.960 --> 00:33:05.880 align:middle line:84% of languages having a bunch of comparatively simple conditions 00:33:05.880 --> 00:33:09.960 align:middle line:84% on what makes a word a well-formed word that interact 00:33:09.960 --> 00:33:12.180 align:middle line:84% with each other to give behaviors 00:33:12.180 --> 00:33:14.220 align:middle line:90% which can be quite complicated. 00:33:14.220 --> 00:33:15.750 align:middle line:84% And this all has to do with stress. 00:33:15.750 --> 00:33:18.000 align:middle line:84% So we've already talked some about stress. 00:33:18.000 --> 00:33:19.720 align:middle line:84% It was part of the flapping rule. 00:33:19.720 --> 00:33:21.660 align:middle line:84% So when we were dealing with how to deal 00:33:21.660 --> 00:33:25.680 align:middle line:84% with the difference in the pronunciation of the "t" 00:33:25.680 --> 00:33:28.710 align:middle line:84% in these two words, "atom" and "atomic," 00:33:28.710 --> 00:33:32.310 align:middle line:84% we said the rule that converts "t" to a flap 00:33:32.310 --> 00:33:34.590 align:middle line:84% needs to be sensitive to where stress 00:33:34.590 --> 00:33:36.900 align:middle line:84% is, so it only happens after stress, 00:33:36.900 --> 00:33:39.000 align:middle line:90% doesn't happen before stress. 00:33:39.000 --> 00:33:42.390 align:middle line:84% And that's why we get it in "atom" but not in "atomic." 00:33:42.390 --> 00:33:45.190 align:middle line:84% Stress often has effects on vowel quality. 00:33:45.190 --> 00:33:47.370 align:middle line:84% So if I were actually writing "atom" and "atomic," 00:33:47.370 --> 00:33:48.990 align:middle line:84% it would look something like that. 00:33:48.990 --> 00:33:54.660 align:middle line:84% You can see that "atom" and "atomic," 00:33:54.660 --> 00:33:57.690 align:middle line:84% if we ignore the "-ic," "atom" and "atomic" 00:33:57.690 --> 00:34:00.930 align:middle line:84% have almost nothing in common apart from an "m." 00:34:00.930 --> 00:34:05.100 align:middle line:84% So the two vowels are different and the sound 00:34:05.100 --> 00:34:08.550 align:middle line:84% that's spelled with a "t" is different in the two words. 00:34:08.550 --> 00:34:10.358 align:middle line:84% I think I referred to things like this 00:34:10.358 --> 00:34:12.150 align:middle line:84% when I was talking about the one good thing 00:34:12.150 --> 00:34:14.730 align:middle line:84% about English spelling, which is that a given morpheme is 00:34:14.730 --> 00:34:17.730 align:middle line:84% typically spelled more or less the same way, even if it's 00:34:17.730 --> 00:34:19.335 align:middle line:90% pronounced quite differently. 00:34:19.335 --> 00:34:23.300 align:middle line:90% 00:34:23.300 --> 00:34:26.630 align:middle line:84% Stress is also important for intonation. 00:34:26.630 --> 00:34:28.820 align:middle line:84% We won't have time to talk seriously 00:34:28.820 --> 00:34:31.130 align:middle line:90% about intonation in this class. 00:34:31.130 --> 00:34:33.530 align:middle line:84% We probably won't have time to talk about it at all. 00:34:33.530 --> 00:34:35.840 align:middle line:84% Intonation is the study of the rules for how 00:34:35.840 --> 00:34:39.080 align:middle line:84% the pitch of your voice rises and falls as you speak, 00:34:39.080 --> 00:34:41.570 align:middle line:84% which, as you can imagine, is a really complicated topic, 00:34:41.570 --> 00:34:43.969 align:middle line:90% and it's really interesting. 00:34:43.969 --> 00:34:49.310 align:middle line:90% One kind of thing for which-- 00:34:49.310 --> 00:34:51.949 align:middle line:84% here's a quick and easy example of stress 00:34:51.949 --> 00:34:53.690 align:middle line:90% being important for intonation. 00:34:53.690 --> 00:34:56.756 align:middle line:90% There's a kind of-- 00:34:56.756 --> 00:34:59.450 align:middle line:90% it's called evocative chant. 00:34:59.450 --> 00:35:01.670 align:middle line:84% It's used in various circumstances, 00:35:01.670 --> 00:35:03.765 align:middle line:90% but one is when you are calling. 00:35:03.765 --> 00:35:05.390 align:middle line:84% I especially think of it as when you're 00:35:05.390 --> 00:35:07.557 align:middle line:84% calling a child-- like you're a parent trying to get 00:35:07.557 --> 00:35:09.710 align:middle line:90% the children to come home. 00:35:09.710 --> 00:35:14.000 align:middle line:84% You say things like-- you say their name and you say 00:35:14.000 --> 00:35:16.040 align:middle line:90% "Lau-ra." 00:35:16.040 --> 00:35:17.495 align:middle line:90% There's a tune to that. 00:35:17.495 --> 00:35:20.360 align:middle line:90% 00:35:20.360 --> 00:35:22.940 align:middle line:84% If I were putting the tune here, you might say yeah, 00:35:22.940 --> 00:35:28.100 align:middle line:84% there's a higher note, "Lau-," and a lower note, "-ra." 00:35:28.100 --> 00:35:30.340 align:middle line:84% That's the way to call Laura's name. 00:35:30.340 --> 00:35:32.090 align:middle line:84% You use that when you're calling children. 00:35:32.090 --> 00:35:34.225 align:middle line:84% You use it when you're selling hot dogs. 00:35:34.225 --> 00:35:35.600 align:middle line:84% So if you're in the ballpark, you 00:35:35.600 --> 00:35:38.330 align:middle line:84% say "hot dogs, popcorn, peanuts." 00:35:38.330 --> 00:35:41.030 align:middle line:84% There's this high-low tune that you sing when 00:35:41.030 --> 00:35:43.940 align:middle line:90% you're doing things like that. 00:35:43.940 --> 00:35:46.470 align:middle line:90% How do you call Pierre? 00:35:46.470 --> 00:35:50.230 align:middle line:90% 00:35:50.230 --> 00:35:52.240 align:middle line:90% "Pierre." 00:35:52.240 --> 00:35:55.330 align:middle line:90% It isn't "Pi-erre," yeah? 00:35:55.330 --> 00:35:58.030 align:middle line:90% It's "Pierre." 00:35:58.030 --> 00:35:59.600 align:middle line:90% So its high goes here. 00:35:59.600 --> 00:36:01.630 align:middle line:90% And then maybe also a low. 00:36:01.630 --> 00:36:04.090 align:middle line:84% And over here there's, I don't know, a low. 00:36:04.090 --> 00:36:06.920 align:middle line:90% "Pierre." 00:36:06.920 --> 00:36:10.380 align:middle line:84% What's the difference between Laura and Pierre? 00:36:10.380 --> 00:36:12.810 align:middle line:84% I mean, Laura is a girl and Pierre is a boy. 00:36:12.810 --> 00:36:15.038 align:middle line:90% But what else? 00:36:15.038 --> 00:36:16.830 align:middle line:84% AUDIENCE: Stress is on the end in "Pierre." 00:36:16.830 --> 00:36:18.720 align:middle line:84% It's on the first syllable with "Laura." 00:36:18.720 --> 00:36:19.970 align:middle line:84% NORVIN RICHARDS: Yeah, so the high-- 00:36:19.970 --> 00:36:21.850 align:middle line:84% we were just talking about how you do the vocative chant. 00:36:21.850 --> 00:36:23.805 align:middle line:84% The high is on the first syllable of Laura. 00:36:23.805 --> 00:36:27.935 align:middle line:84% The high is on the second syllable of Pierre. 00:36:27.935 --> 00:36:28.435 align:middle line:90% Why? 00:36:28.435 --> 00:36:31.440 align:middle line:90% 00:36:31.440 --> 00:36:37.130 align:middle line:84% Let's chant some other names, longer names. 00:36:37.130 --> 00:36:43.670 align:middle line:90% 00:36:43.670 --> 00:36:45.350 align:middle line:90% Genevieve. 00:36:45.350 --> 00:36:46.417 align:middle line:90% "Genevieve." 00:36:46.417 --> 00:36:47.000 align:middle line:90% Is that right? 00:36:47.000 --> 00:36:51.500 align:middle line:90% "Genevieve," I think. 00:36:51.500 --> 00:36:54.320 align:middle line:84% Maybe this is the low here at the end. 00:36:54.320 --> 00:36:56.020 align:middle line:90% "Genevieve," hi ho. 00:36:56.020 --> 00:36:57.710 align:middle line:90% Yeah. 00:36:57.710 --> 00:37:03.290 align:middle line:84% If you're selling vegetables at the ballpark 00:37:03.290 --> 00:37:04.295 align:middle line:90% along with hot dogs-- 00:37:04.295 --> 00:37:07.830 align:middle line:90% 00:37:07.830 --> 00:37:08.850 align:middle line:90% AUDIENCE: "Asparagus." 00:37:08.850 --> 00:37:10.470 align:middle line:84% NORVIN RICHARDS: "Asparagus," I think. 00:37:10.470 --> 00:37:14.460 align:middle line:90% 00:37:14.460 --> 00:37:15.420 align:middle line:90% This is low. 00:37:15.420 --> 00:37:25.948 align:middle line:90% 00:37:25.948 --> 00:37:26.865 align:middle line:90% AUDIENCE: "Cucumbers." 00:37:26.865 --> 00:37:27.330 align:middle line:90% NORVIN RICHARDS: "Cu--" 00:37:27.330 --> 00:37:28.260 align:middle line:90% AUDIENCE: "-cumbers." 00:37:28.260 --> 00:37:30.450 align:middle line:90% NORVIN RICHARDS: "Cucumbers." 00:37:30.450 --> 00:37:33.180 align:middle line:84% Think is it-- wait, is it "cucumbers?" 00:37:33.180 --> 00:37:34.533 align:middle line:90% Or is it "cucumbers"? 00:37:34.533 --> 00:37:35.825 align:middle line:90% AUDIENCE: No, that sounds good. 00:37:35.825 --> 00:37:37.785 align:middle line:90% Yeah. 00:37:37.785 --> 00:37:38.910 align:middle line:90% NORVIN RICHARDS: Cucumbers. 00:37:38.910 --> 00:37:43.380 align:middle line:84% Yeah, so this is high and this is low, Yeah? 00:37:43.380 --> 00:37:46.270 align:middle line:90% 00:37:46.270 --> 00:37:48.270 align:middle line:90% Let me do one more. 00:37:48.270 --> 00:37:49.890 align:middle line:84% Suppose you're selling jalapenos. 00:37:49.890 --> 00:37:54.963 align:middle line:90% 00:37:54.963 --> 00:37:55.880 align:middle line:90% AUDIENCE: "Jalapenos." 00:37:55.880 --> 00:37:57.470 align:middle line:90% NORVIN RICHARDS: "Jalapenos." 00:37:57.470 --> 00:37:58.205 align:middle line:90% Here's the high. 00:37:58.205 --> 00:38:01.352 align:middle line:90% 00:38:01.352 --> 00:38:02.810 align:middle line:84% Forget about the lows for a second. 00:38:02.810 --> 00:38:06.440 align:middle line:90% 00:38:06.440 --> 00:38:09.480 align:middle line:90% Where does the high go? 00:38:09.480 --> 00:38:09.980 align:middle line:90% Yes. 00:38:09.980 --> 00:38:11.210 align:middle line:90% AUDIENCE: The stressed syllable. 00:38:11.210 --> 00:38:13.085 align:middle line:84% NORVIN RICHARDS: The stressed syllable, yeah? 00:38:13.085 --> 00:38:18.410 align:middle line:84% So it goes on "Genevieve," "Laura," "Pierre," yeah. 00:38:18.410 --> 00:38:22.070 align:middle line:84% "Cucumbers," "asparagus," and "jalapenos," that's 00:38:22.070 --> 00:38:23.972 align:middle line:90% where the high starts, yeah. 00:38:23.972 --> 00:38:25.430 align:middle line:84% And before that high you have lows. 00:38:25.430 --> 00:38:25.880 align:middle line:90% Yeah. 00:38:25.880 --> 00:38:27.410 align:middle line:84% AUDIENCE: There are some situations where we keep 00:38:27.410 --> 00:38:28.550 align:middle line:90% the high for a few syllables. 00:38:28.550 --> 00:38:29.090 align:middle line:90% NORVIN RICHARDS: Yeah. 00:38:29.090 --> 00:38:30.410 align:middle line:90% AUDIENCE: After the "s." 00:38:30.410 --> 00:38:32.740 align:middle line:84% NORVIN RICHARDS: Yeah, like where? 00:38:32.740 --> 00:38:34.742 align:middle line:84% AUDIENCE: "Cu"-- wait, oh, "Genevieve." 00:38:34.742 --> 00:38:35.950 align:middle line:90% NORVIN RICHARDS: "Genevieve." 00:38:35.950 --> 00:38:37.100 align:middle line:90% AUDIENCE: And "asparagus." 00:38:37.100 --> 00:38:38.780 align:middle line:84% NORVIN RICHARDS: And "asparagus." 00:38:38.780 --> 00:38:40.220 align:middle line:90% No, "asparagus." 00:38:40.220 --> 00:38:42.740 align:middle line:84% Yeah, so for those you've just get an "l" at the end 00:38:42.740 --> 00:38:44.330 align:middle line:90% and it's high until that. 00:38:44.330 --> 00:38:47.000 align:middle line:84% Maybe in all of these, you get an "l" at the end, right? 00:38:47.000 --> 00:38:49.580 align:middle line:84% Even in Pierre, where there's a high at the end, 00:38:49.580 --> 00:38:50.570 align:middle line:90% there's also a low. 00:38:50.570 --> 00:38:52.910 align:middle line:84% I think you don't say Pierre, right? 00:38:52.910 --> 00:38:54.410 align:middle line:90% It's Pierre. 00:38:54.410 --> 00:38:56.270 align:middle line:90% You have to go back down to low. 00:38:56.270 --> 00:39:00.260 align:middle line:84% So the vocative chant goes put a high on the stressed syllable, 00:39:00.260 --> 00:39:02.600 align:middle line:90% put a low at the end, yeah? 00:39:02.600 --> 00:39:06.680 align:middle line:84% And then what's the difference between "cucumbers" 00:39:06.680 --> 00:39:07.790 align:middle line:90% and "asparagus"? 00:39:07.790 --> 00:39:12.520 align:middle line:90% 00:39:12.520 --> 00:39:18.080 align:middle line:84% Or between "Genevieve"-- wait, do we want "Genevieve"? 00:39:18.080 --> 00:39:19.330 align:middle line:90% No, we don't want "Genevieve." 00:39:19.330 --> 00:39:21.170 align:middle line:90% Sorry, Genevieve. 00:39:21.170 --> 00:39:22.420 align:middle line:90% My sister's name is Genevieve. 00:39:22.420 --> 00:39:23.795 align:middle line:84% Nobody tell her that I said that. 00:39:23.795 --> 00:39:27.500 align:middle line:90% 00:39:27.500 --> 00:39:30.190 align:middle line:84% What's the difference between "cucumbers" and "asparagus"? 00:39:30.190 --> 00:39:31.990 align:middle line:90% Yeah. 00:39:31.990 --> 00:39:36.071 align:middle line:84% AUDIENCE: Well, we haven't learned syllable division yet. 00:39:36.071 --> 00:39:36.988 align:middle line:90% NORVIN RICHARDS: Yeah. 00:39:36.988 --> 00:39:39.610 align:middle line:84% AUDIENCE: Singing "cucumbers" added 00:39:39.610 --> 00:39:41.548 align:middle line:90% part of that second syllable. 00:39:41.548 --> 00:39:43.590 align:middle line:84% NORVIN RICHARDS: Yep, yep, that's probably right. 00:39:43.590 --> 00:39:46.320 align:middle line:84% AUDIENCE: Whereas in "Genevieve" that is not 00:39:46.320 --> 00:39:47.727 align:middle line:90% part of the second syllable. 00:39:47.727 --> 00:39:48.644 align:middle line:90% Same with "asparagus." 00:39:48.644 --> 00:39:49.644 align:middle line:90% NORVIN RICHARDS: Mm-hmm. 00:39:49.644 --> 00:39:50.670 align:middle line:90% Yeah. 00:39:50.670 --> 00:39:52.080 align:middle line:90% Yeah, that's probably right. 00:39:52.080 --> 00:39:54.180 align:middle line:84% Yeah, yeah, that's probably true. 00:39:54.180 --> 00:39:55.095 align:middle line:90% Does it? 00:39:55.095 --> 00:39:56.720 align:middle line:84% AUDIENCE: We have some data like you're 00:39:56.720 --> 00:39:58.790 align:middle line:90% high until you hit an obstruent. 00:39:58.790 --> 00:39:59.440 align:middle line:90% And then-- 00:39:59.440 --> 00:40:00.732 align:middle line:90% NORVIN RICHARDS: Ooh, oh, gosh. 00:40:00.732 --> 00:40:03.050 align:middle line:90% 00:40:03.050 --> 00:40:05.057 align:middle line:84% You're high until you hit an obstruent. 00:40:05.057 --> 00:40:06.515 align:middle line:84% AUDIENCE: So low until they stress. 00:40:06.515 --> 00:40:07.340 align:middle line:90% And then-- 00:40:07.340 --> 00:40:08.950 align:middle line:90% NORVIN RICHARDS: Oh, man. 00:40:08.950 --> 00:40:10.700 align:middle line:84% Let me try to think of some more examples. 00:40:10.700 --> 00:40:12.830 align:middle line:84% That can't possibly be right, but it covers the data 00:40:12.830 --> 00:40:13.747 align:middle line:90% that we've got so far. 00:40:13.747 --> 00:40:15.885 align:middle line:90% Yes. 00:40:15.885 --> 00:40:16.385 align:middle line:90% [LAUGHS] 00:40:16.385 --> 00:40:16.820 align:middle line:90% AUDIENCE: The last one. 00:40:16.820 --> 00:40:18.080 align:middle line:84% NORVIN RICHARDS: The last one has to be low, yeah. 00:40:18.080 --> 00:40:19.430 align:middle line:90% So the last one has to be low. 00:40:19.430 --> 00:40:20.120 align:middle line:90% That's right. 00:40:20.120 --> 00:40:22.760 align:middle line:84% And the high goes on the stressed syllable. 00:40:22.760 --> 00:40:26.725 align:middle line:84% And then you stay high maybe, so far. 00:40:26.725 --> 00:40:28.100 align:middle line:84% For most of these what happens is 00:40:28.100 --> 00:40:30.350 align:middle line:84% that you stay high until you hit a low, right? 00:40:30.350 --> 00:40:31.640 align:middle line:90% So it's "Genevieve." 00:40:31.640 --> 00:40:34.340 align:middle line:84% The "vieve" is at the end, or it's "asparagus" 00:40:34.340 --> 00:40:35.610 align:middle line:90% or it's "jalapenos." 00:40:35.610 --> 00:40:38.810 align:middle line:90% 00:40:38.810 --> 00:40:40.410 align:middle line:90% And these are low. 00:40:40.410 --> 00:40:42.372 align:middle line:84% So you're low before you get to the high. 00:40:42.372 --> 00:40:44.330 align:middle line:84% And the real question is, what's the difference 00:40:44.330 --> 00:40:46.310 align:middle line:84% between "cucumbers" and "asparagus," 00:40:46.310 --> 00:40:49.220 align:middle line:84% where this high spreads and this high just stays here? 00:40:49.220 --> 00:40:52.070 align:middle line:84% Why is this low not just-- why is there a low not just 00:40:52.070 --> 00:40:54.110 align:middle line:90% here, but also here? 00:40:54.110 --> 00:40:55.550 align:middle line:90% And I think that's true, right? 00:40:55.550 --> 00:40:57.800 align:middle line:84% We agreed that it is "cucumbers." 00:40:57.800 --> 00:40:59.060 align:middle line:90% It isn't "cucumbers." 00:40:59.060 --> 00:41:04.710 align:middle line:90% 00:41:04.710 --> 00:41:07.074 align:middle line:84% Let me try to think of another word. 00:41:07.074 --> 00:41:08.370 align:middle line:90% AUDIENCE: "Endocrinology." 00:41:08.370 --> 00:41:09.750 align:middle line:84% NORVIN RICHARDS: "Endocrinology." 00:41:09.750 --> 00:41:14.970 align:middle line:90% 00:41:14.970 --> 00:41:15.870 align:middle line:90% What's a better word? 00:41:15.870 --> 00:41:16.710 align:middle line:90% That's a good word. 00:41:16.710 --> 00:41:18.703 align:middle line:90% I'm sorry. 00:41:18.703 --> 00:41:20.370 align:middle line:84% I'm trying to think of another word that 00:41:20.370 --> 00:41:22.635 align:middle line:84% has the property that I'm looking for. 00:41:22.635 --> 00:41:25.870 align:middle line:90% 00:41:25.870 --> 00:41:28.630 align:middle line:90% Lend a, "interdisciplinary." 00:41:28.630 --> 00:41:30.130 align:middle line:90% "Interdisciplinary. 00:41:30.130 --> 00:41:31.130 align:middle line:90% Interdisciplinary. 00:41:31.130 --> 00:41:33.750 align:middle line:90% Interdisciplinary." 00:41:33.750 --> 00:41:36.000 align:middle line:90% Yeah, no. 00:41:36.000 --> 00:41:37.080 align:middle line:90% "Interdisciplinary." 00:41:37.080 --> 00:41:39.060 align:middle line:84% So suppose I'm selling interdisciplinary. 00:41:39.060 --> 00:41:40.410 align:middle line:90% [LAUGHTER] 00:41:40.410 --> 00:41:43.950 align:middle line:84% Or suppose I've named my child "Interdisciplinary." 00:41:43.950 --> 00:41:48.010 align:middle line:84% It's "Interdisciplinary," is that right? 00:41:48.010 --> 00:41:51.660 align:middle line:90% Yeah, I think that's right. 00:41:51.660 --> 00:41:56.430 align:middle line:84% So my-- that's a nice name for a child. 00:41:56.430 --> 00:42:02.430 align:middle line:90% I mean, "Interdisciplinary." 00:42:02.430 --> 00:42:04.290 align:middle line:90% It's low here, yeah? 00:42:04.290 --> 00:42:08.140 align:middle line:90% 00:42:08.140 --> 00:42:12.280 align:middle line:84% Man, so let me tell you what I think is going on. 00:42:12.280 --> 00:42:14.530 align:middle line:84% We've talked about words as though they only 00:42:14.530 --> 00:42:20.080 align:middle line:84% have one stress, "cucumber," "asparagus," "Genevieve." 00:42:20.080 --> 00:42:22.420 align:middle line:90% Let's take "interdisciplinary." 00:42:22.420 --> 00:42:24.700 align:middle line:84% How many stresses are there in "interdisciplinary"? 00:42:24.700 --> 00:42:26.860 align:middle line:84% Where's the main stress in "interdisciplinary"? 00:42:26.860 --> 00:42:28.120 align:middle line:90% AUDIENCE: It's "dis" here. 00:42:28.120 --> 00:42:29.120 align:middle line:84% NORVIN RICHARDS: Yeah, it's "dis," 00:42:29.120 --> 00:42:30.460 align:middle line:90% it's here, where the high is. 00:42:30.460 --> 00:42:31.960 align:middle line:84% So the high goes on the main stress. 00:42:31.960 --> 00:42:34.510 align:middle line:84% What other syllables in this are stressed? 00:42:34.510 --> 00:42:35.268 align:middle line:90% AUDIENCE: "-nary." 00:42:35.268 --> 00:42:36.310 align:middle line:90% NORVIN RICHARDS: "-nary." 00:42:36.310 --> 00:42:37.930 align:middle line:90% There's another stress. 00:42:37.930 --> 00:42:40.030 align:middle line:90% And maybe here, I don't know. 00:42:40.030 --> 00:42:41.530 align:middle line:90% AUDIENCE: Really, OK. 00:42:41.530 --> 00:42:43.348 align:middle line:90% NORVIN RICHARDS: Yeah. 00:42:43.348 --> 00:42:45.550 align:middle line:84% AUDIENCE: Is this high on the main stress 00:42:45.550 --> 00:42:47.320 align:middle line:84% and then it stays high until it reaches 00:42:47.320 --> 00:42:48.730 align:middle line:90% the next like substress? 00:42:48.730 --> 00:42:50.120 align:middle line:84% NORVIN RICHARDS: I think it's something like that. 00:42:50.120 --> 00:42:51.880 align:middle line:84% So it's high and it stays high until you get 00:42:51.880 --> 00:42:53.140 align:middle line:90% to the next stressed syllable. 00:42:53.140 --> 00:42:55.460 align:middle line:84% The next stressed syllable is where the low starts. 00:42:55.460 --> 00:42:55.960 align:middle line:90% Yeah. 00:42:55.960 --> 00:42:58.880 align:middle line:84% And there's always a low at the end. 00:42:58.880 --> 00:43:01.840 align:middle line:84% So for the difference between "cucumbers" and "asparagus," 00:43:01.840 --> 00:43:04.450 align:middle line:84% I think, is that "cucumber" has a secondary stress 00:43:04.450 --> 00:43:08.560 align:middle line:84% on the second syllable, whereas "asparagus" only has 00:43:08.560 --> 00:43:09.100 align:middle line:90% one stress. 00:43:09.100 --> 00:43:09.693 align:middle line:90% It's here. 00:43:09.693 --> 00:43:11.860 align:middle line:84% Yeah, and that might be related to your observation. 00:43:11.860 --> 00:43:14.830 align:middle line:84% This has a closed syllable here and closed syllables are often 00:43:14.830 --> 00:43:17.450 align:middle line:90% associated with stress. 00:43:17.450 --> 00:43:18.980 align:middle line:84% This will not be on the final exam. 00:43:18.980 --> 00:43:20.000 align:middle line:90% There is no final exam. 00:43:20.000 --> 00:43:20.500 align:middle line:90% Yeah. 00:43:20.500 --> 00:43:24.190 align:middle line:84% But here's an example of stress being useful 00:43:24.190 --> 00:43:27.070 align:middle line:84% for understanding a phenomenon, in this case, the phenomenon 00:43:27.070 --> 00:43:29.283 align:middle line:90% of the vocative chant. 00:43:29.283 --> 00:43:30.700 align:middle line:84% All of you are about to be invited 00:43:30.700 --> 00:43:34.120 align:middle line:84% to come up with a research topic, something to work on. 00:43:34.120 --> 00:43:37.840 align:middle line:84% This is something you could try working on with the person 00:43:37.840 --> 00:43:39.430 align:middle line:90% that you're working with. 00:43:39.430 --> 00:43:41.860 align:middle line:84% Ask them how they call their children, right? 00:43:41.860 --> 00:43:43.450 align:middle line:90% And also how they sell things. 00:43:43.450 --> 00:43:45.400 align:middle line:84% And find out whether there's a special tune 00:43:45.400 --> 00:43:48.910 align:middle line:84% for that, because in many languages there is. 00:43:48.910 --> 00:43:51.560 align:middle line:84% It would be interesting to see, yeah. 00:43:51.560 --> 00:43:53.200 align:middle line:90% OK, all right. 00:43:53.200 --> 00:43:56.830 align:middle line:90% 00:43:56.830 --> 00:43:59.092 align:middle line:84% So, yeah, this was all an attempt 00:43:59.092 --> 00:44:01.300 align:middle line:84% to convince you that it might be interesting to think 00:44:01.300 --> 00:44:02.110 align:middle line:90% about stress. 00:44:02.110 --> 00:44:05.020 align:middle line:84% It's relevant for various types of phonological phenomenon. 00:44:05.020 --> 00:44:07.930 align:middle line:84% It was relevant for the flapping rule in English. 00:44:07.930 --> 00:44:11.110 align:middle line:84% We've just seen it's relevant for the rules 00:44:11.110 --> 00:44:12.130 align:middle line:90% for vocative chant. 00:44:12.130 --> 00:44:15.128 align:middle line:90% 00:44:15.128 --> 00:44:16.670 align:middle line:84% I'm going to leave those on the board 00:44:16.670 --> 00:44:18.560 align:middle line:84% just so future classes can wonder 00:44:18.560 --> 00:44:21.290 align:middle line:90% what the heck we were doing. 00:44:21.290 --> 00:44:26.228 align:middle line:84% In languages where words have one fixed position for stress-- 00:44:26.228 --> 00:44:28.770 align:middle line:84% so there are languages out there where the position of stress 00:44:28.770 --> 00:44:30.390 align:middle line:90% is always fixed. 00:44:30.390 --> 00:44:32.030 align:middle line:84% These are data from a database called 00:44:32.030 --> 00:44:36.200 align:middle line:84% WALS, the world atlas of language structures, 00:44:36.200 --> 00:44:38.930 align:middle line:90% which is free and accessible. 00:44:38.930 --> 00:44:42.680 align:middle line:84% I recommend going and having a look at it. 00:44:42.680 --> 00:44:46.970 align:middle line:84% When I got these data, there were 282 languages 00:44:46.970 --> 00:44:48.620 align:middle line:90% described in the WALS database. 00:44:48.620 --> 00:44:50.900 align:middle line:84% The WALS database describes languages with respect 00:44:50.900 --> 00:44:55.040 align:middle line:84% to a variety of phenomena, not always accurately, I 00:44:55.040 --> 00:44:55.650 align:middle line:90% have to say. 00:44:55.650 --> 00:44:58.280 align:middle line:84% So if you have a particular language you're interested in, 00:44:58.280 --> 00:44:59.280 align:middle line:90% you should double check. 00:44:59.280 --> 00:45:03.560 align:middle line:84% But it's a nice first pass, looking at a bunch of data. 00:45:03.560 --> 00:45:08.270 align:middle line:84% In that database, you can see there are, among languages 00:45:08.270 --> 00:45:12.230 align:middle line:84% that have just one place where stress always goes, 00:45:12.230 --> 00:45:15.143 align:middle line:84% there are languages where that one 00:45:15.143 --> 00:45:16.310 align:middle line:90% place is the first syllable. 00:45:16.310 --> 00:45:17.450 align:middle line:90% That's pretty common. 00:45:17.450 --> 00:45:20.970 align:middle line:84% About a third of the languages like that, that do that. 00:45:20.970 --> 00:45:25.460 align:middle line:84% The most common type of stress for a language like that 00:45:25.460 --> 00:45:26.780 align:middle line:90% is penultimate stress. 00:45:26.780 --> 00:45:29.570 align:middle line:84% Languages like Zulu and Polish have stress on the next 00:45:29.570 --> 00:45:31.670 align:middle line:90% to last syllable. 00:45:31.670 --> 00:45:34.430 align:middle line:84% And then you can see there are some other patterns that 00:45:34.430 --> 00:45:35.250 align:middle line:90% are less common. 00:45:35.250 --> 00:45:38.940 align:middle line:84% So the last syllable is not uncommon. 00:45:38.940 --> 00:45:41.270 align:middle line:84% And then there are some really weird systems 00:45:41.270 --> 00:45:44.630 align:middle line:84% like Hocak, which is a Native American language, 00:45:44.630 --> 00:45:46.760 align:middle line:84% has stress on its third syllable. 00:45:46.760 --> 00:45:47.662 align:middle line:90% That's quite rare. 00:45:47.662 --> 00:45:49.370 align:middle line:84% There are some languages that have stress 00:45:49.370 --> 00:45:54.952 align:middle line:84% on the antepenultimate syllable, antepenultimate, yeah, 00:45:54.952 --> 00:45:57.410 align:middle line:84% and some languages that have stress on the second syllable. 00:45:57.410 --> 00:45:59.713 align:middle line:84% So these are all patterns that you find. 00:45:59.713 --> 00:46:01.130 align:middle line:84% As you can see, these are patterns 00:46:01.130 --> 00:46:06.590 align:middle line:84% that are best described as stressing a syllable that 00:46:06.590 --> 00:46:12.710 align:middle line:84% is counted either from the left edge or the right of the word. 00:46:12.710 --> 00:46:16.070 align:middle line:84% There are other systems that are not hard to imagine 00:46:16.070 --> 00:46:17.340 align:middle line:90% but that don't exist. 00:46:17.340 --> 00:46:21.830 align:middle line:84% And so people who are developing theories of stress typology, 00:46:21.830 --> 00:46:23.690 align:middle line:84% one of their goals, for example, is 00:46:23.690 --> 00:46:26.690 align:middle line:84% to have a theory that rules out languages with the rule, 00:46:26.690 --> 00:46:28.220 align:middle line:84% stress the syllable that's closest 00:46:28.220 --> 00:46:30.020 align:middle line:90% to the middle of the word. 00:46:30.020 --> 00:46:32.210 align:middle line:84% It's distressingly easy to come up 00:46:32.210 --> 00:46:34.520 align:middle line:84% with a theory that predicts the existence of languages 00:46:34.520 --> 00:46:35.030 align:middle line:90% like that. 00:46:35.030 --> 00:46:36.720 align:middle line:90% And they don't exist. 00:46:36.720 --> 00:46:42.900 align:middle line:84% So you want your theory to not capture languages like that. 00:46:42.900 --> 00:46:46.760 align:middle line:84% One way to describe this kind of pattern is to have constraints, 00:46:46.760 --> 00:46:48.740 align:middle line:84% you know, like the constraints that say things 00:46:48.740 --> 00:46:51.650 align:middle line:84% like don't have three consonants in a row, constraints that 00:46:51.650 --> 00:46:55.190 align:middle line:84% say things like put stress as far to the left or as far 00:46:55.190 --> 00:46:57.690 align:middle line:84% to the right in the word as possible. 00:46:57.690 --> 00:47:02.840 align:middle line:84% And then have that constraint interact with other constraints 00:47:02.840 --> 00:47:05.480 align:middle line:84% that say things like don't stress the last syllable 00:47:05.480 --> 00:47:08.180 align:middle line:84% or don't stress the first syllable. 00:47:08.180 --> 00:47:10.950 align:middle line:84% So for example, the quite common pattern 00:47:10.950 --> 00:47:12.950 align:middle line:84% that you get in Zulu and Polish where you stress 00:47:12.950 --> 00:47:15.170 align:middle line:90% the next to last syllable. 00:47:15.170 --> 00:47:17.540 align:middle line:84% That's one that says put stress as far 00:47:17.540 --> 00:47:19.340 align:middle line:84% to the right in the word as you can, oh, 00:47:19.340 --> 00:47:21.262 align:middle line:84% but don't stress the last syllable. 00:47:21.262 --> 00:47:22.970 align:middle line:84% So don't stress the last syllable, that's 00:47:22.970 --> 00:47:24.080 align:middle line:90% the most important thing. 00:47:24.080 --> 00:47:27.570 align:middle line:84% But putting aside that, put stress as far to the right 00:47:27.570 --> 00:47:28.070 align:middle line:90% as you can. 00:47:28.070 --> 00:47:30.477 align:middle line:84% That means put stress on the next-to-last syllable. 00:47:30.477 --> 00:47:32.810 align:middle line:84% That's one way of describing that kind of stress system. 00:47:32.810 --> 00:47:33.530 align:middle line:90% Yeah. 00:47:33.530 --> 00:47:34.160 align:middle line:90% Yes. 00:47:34.160 --> 00:47:35.530 align:middle line:84% AUDIENCE: Are these languages sensitive to 00:47:35.530 --> 00:47:37.488 align:middle line:84% deal with a verb which comes in the last place, 00:47:37.488 --> 00:47:39.160 align:middle line:84% anyway it felt like it was for that. 00:47:39.160 --> 00:47:40.900 align:middle line:84% NORVIN RICHARDS: Yeah, so there are languages-- 00:47:40.900 --> 00:47:42.970 align:middle line:84% we looked at this in Lardil, that just don't like 00:47:42.970 --> 00:47:44.140 align:middle line:90% words of only one syllable. 00:47:44.140 --> 00:47:46.093 align:middle line:90% They rule them out. 00:47:46.093 --> 00:47:47.260 align:middle line:90% But you're absolutely right. 00:47:47.260 --> 00:47:52.060 align:middle line:84% So languages that have systems like these, typically it's like 00:47:52.060 --> 00:47:55.250 align:middle line:84% stress the next-to-last syllable if there is one. 00:47:55.250 --> 00:47:58.510 align:middle line:84% So in one monosyllabic word, you stress the one syllable 00:47:58.510 --> 00:47:59.210 align:middle line:90% that you've got. 00:47:59.210 --> 00:48:01.060 align:middle line:90% Yeah, that's a very good point. 00:48:01.060 --> 00:48:04.140 align:middle line:84% That's a thing that you find in these languages. 00:48:04.140 --> 00:48:08.190 align:middle line:90% Yeah, yeah, yeah. 00:48:08.190 --> 00:48:14.190 align:middle line:84% It would be nice if there were some correlation between having 00:48:14.190 --> 00:48:19.950 align:middle line:84% a stress system that wanted to avoid final stress, let's say, 00:48:19.950 --> 00:48:25.143 align:middle line:84% and being like Lardil in not liking monosyllables, right? 00:48:25.143 --> 00:48:27.060 align:middle line:84% It would be nice if the world worked that way. 00:48:27.060 --> 00:48:28.620 align:middle line:90% As far as I know it doesn't. 00:48:28.620 --> 00:48:31.695 align:middle line:84% So it doesn't seem as though the condition that says it 00:48:31.695 --> 00:48:33.570 align:middle line:84% would be better for your words to be at least 00:48:33.570 --> 00:48:36.240 align:middle line:84% two syllables long, it doesn't seem to be motivated 00:48:36.240 --> 00:48:37.768 align:middle line:90% by these considerations. 00:48:37.768 --> 00:48:39.810 align:middle line:84% I sure wish that it was, because it would be nice 00:48:39.810 --> 00:48:42.960 align:middle line:84% if there-- then we would have an explanation for part of that. 00:48:42.960 --> 00:48:45.240 align:middle line:84% But it's not clear to me that there is any connection 00:48:45.240 --> 00:48:48.170 align:middle line:90% between those things. 00:48:48.170 --> 00:48:48.670 align:middle line:90% Yeah. 00:48:48.670 --> 00:48:51.600 align:middle line:90% 00:48:51.600 --> 00:48:55.050 align:middle line:84% Now those, we were just talking about languages in which stress 00:48:55.050 --> 00:48:56.490 align:middle line:90% is only in one place. 00:48:56.490 --> 00:48:58.590 align:middle line:84% But as we actually saw as we were looking 00:48:58.590 --> 00:49:00.758 align:middle line:84% at the vocative chant, it's possible for a language 00:49:00.758 --> 00:49:02.550 align:middle line:84% to have multiple stresses in a single word. 00:49:02.550 --> 00:49:04.570 align:middle line:90% English does that all the time. 00:49:04.570 --> 00:49:06.660 align:middle line:84% And there are languages out there 00:49:06.660 --> 00:49:11.370 align:middle line:84% that have patterns where you stress more than one syllable. 00:49:11.370 --> 00:49:13.440 align:middle line:84% Here are some examples from Pintupi, 00:49:13.440 --> 00:49:16.350 align:middle line:84% which is an Aboriginal language of Australia, 00:49:16.350 --> 00:49:20.490 align:middle line:84% in which the basic rule is stress 00:49:20.490 --> 00:49:22.050 align:middle line:90% the odd-numbered syllables. 00:49:22.050 --> 00:49:25.380 align:middle line:84% So stress, put stress on the first, third, fifth, seventh, 00:49:25.380 --> 00:49:29.340 align:middle line:84% and et cetera, syllables of the word. 00:49:29.340 --> 00:49:33.140 align:middle line:84% So again, you can think of that as a little pile 00:49:33.140 --> 00:49:34.070 align:middle line:90% of constraints. 00:49:34.070 --> 00:49:37.490 align:middle line:84% Stress the first syllable, don't stress the last syllable, 00:49:37.490 --> 00:49:41.030 align:middle line:84% and don't allow yourself to have two syllables in a row that 00:49:41.030 --> 00:49:46.130 align:middle line:84% are the same, that is are both stressed or both unstressed. 00:49:46.130 --> 00:49:49.340 align:middle line:84% So, again, you don't always succeed 00:49:49.340 --> 00:49:51.170 align:middle line:90% in doing all of these things. 00:49:51.170 --> 00:49:54.710 align:middle line:84% In a language like this, whenever you have 00:49:54.710 --> 00:49:58.820 align:middle line:84% an odd-numbered of syllables in your word, 00:49:58.820 --> 00:50:01.190 align:middle line:90% you can achieve the first two-- 00:50:01.190 --> 00:50:04.940 align:middle line:84% stress the first syllable, don't stress the last syllable-- 00:50:04.940 --> 00:50:06.980 align:middle line:84% and actually the next one, don't allow two 00:50:06.980 --> 00:50:08.990 align:middle line:90% stressed syllables in a row. 00:50:08.990 --> 00:50:11.420 align:middle line:84% But the last one, don't allow two unstressed syllables 00:50:11.420 --> 00:50:16.100 align:middle line:84% in a row, you achieve that in the words that have 00:50:16.100 --> 00:50:17.367 align:middle line:90% an even number of syllables. 00:50:17.367 --> 00:50:18.950 align:middle line:84% No, yeah, an even number of syllables, 00:50:18.950 --> 00:50:20.750 align:middle line:84% but not the ones that have an odd number of syllables. 00:50:20.750 --> 00:50:21.920 align:middle line:84% If there's an odd number of syllables, 00:50:21.920 --> 00:50:24.212 align:middle line:84% like in the second word there that has three syllables, 00:50:24.212 --> 00:50:25.700 align:middle line:84% you only stress the first syllable. 00:50:25.700 --> 00:50:27.350 align:middle line:84% And you put up with the fact that it 00:50:27.350 --> 00:50:29.000 align:middle line:84% ends in two unstressed syllables, 00:50:29.000 --> 00:50:33.080 align:middle line:84% because there's no way to avoid that while also satisfying 00:50:33.080 --> 00:50:34.670 align:middle line:84% these higher-ranking constraints like 00:50:34.670 --> 00:50:36.873 align:middle line:84% don't stress the last syllable and don't have two 00:50:36.873 --> 00:50:38.040 align:middle line:90% stressed syllables in a row. 00:50:38.040 --> 00:50:38.900 align:middle line:90% Yeah. 00:50:38.900 --> 00:50:42.080 align:middle line:84% AUDIENCE: How come we go through these four different rules, 00:50:42.080 --> 00:50:46.142 align:middle line:84% as opposed to just declaring stress every other syllable? 00:50:46.142 --> 00:50:48.350 align:middle line:84% NORVIN RICHARDS: Stress every other syllable starting 00:50:48.350 --> 00:50:50.240 align:middle line:90% from the left edge, yeah. 00:50:50.240 --> 00:50:51.860 align:middle line:84% Why am I trying to decompose this 00:50:51.860 --> 00:50:53.990 align:middle line:90% into constraints like this? 00:50:53.990 --> 00:50:57.920 align:middle line:84% I guess in part it's an attempt to constrain 00:50:57.920 --> 00:51:01.470 align:middle line:84% the kinds of syllable, of stress systems that we could find, 00:51:01.470 --> 00:51:01.970 align:middle line:90% right? 00:51:01.970 --> 00:51:04.910 align:middle line:84% So if we let ourselves say things like stress 00:51:04.910 --> 00:51:06.655 align:middle line:90% the odd-numbered syllables-- 00:51:06.655 --> 00:51:08.780 align:middle line:84% I mean first, stress the odd-number of syllables by 00:51:08.780 --> 00:51:10.610 align:middle line:84% itself doesn't actually cover this, 00:51:10.610 --> 00:51:12.290 align:middle line:90% even though it's what I said. 00:51:12.290 --> 00:51:15.350 align:middle line:84% So it's really stress the odd-numbered syllables 00:51:15.350 --> 00:51:17.480 align:middle line:84% unless you're going to stress the last syllable. 00:51:17.480 --> 00:51:18.800 align:middle line:90% Don't stress the last syllable. 00:51:18.800 --> 00:51:22.610 align:middle line:84% So even just saying that, what I just said 00:51:22.610 --> 00:51:25.640 align:middle line:84% is best stated in terms of conflicting constraints. 00:51:25.640 --> 00:51:27.318 align:middle line:84% It's like the most important thing is 00:51:27.318 --> 00:51:28.610 align:middle line:90% don't stress the last syllable. 00:51:28.610 --> 00:51:30.440 align:middle line:84% And then underneath that there's stress 00:51:30.440 --> 00:51:32.300 align:middle line:90% the odd-numbered syllables. 00:51:32.300 --> 00:51:35.990 align:middle line:84% And then, by decomposing stress the odd-numbered syllables 00:51:35.990 --> 00:51:39.098 align:middle line:84% into stress the first syllable, don't 00:51:39.098 --> 00:51:40.640 align:middle line:84% have two stressed syllables in a row, 00:51:40.640 --> 00:51:43.460 align:middle line:84% don't have two unstressed syllables in a row, 00:51:43.460 --> 00:51:47.360 align:middle line:84% we constrain the kinds of stress systems that we could find. 00:51:47.360 --> 00:51:49.370 align:middle line:84% So there aren't stress systems that 00:51:49.370 --> 00:51:52.760 align:middle line:84% say stress the prime numbered syllables 00:51:52.760 --> 00:51:55.827 align:middle line:84% or stress the powers of 2, right? 00:51:55.827 --> 00:51:57.410 align:middle line:84% There are certain kinds of constraints 00:51:57.410 --> 00:51:59.450 align:middle line:84% that you use for stress systems and not others. 00:51:59.450 --> 00:52:02.030 align:middle line:84% And it's one of the kinds of things we're trying to get. 00:52:02.030 --> 00:52:03.373 align:middle line:90% A good question. 00:52:03.373 --> 00:52:04.040 align:middle line:90% Other questions? 00:52:04.040 --> 00:52:05.330 align:middle line:90% Yeah. 00:52:05.330 --> 00:52:08.820 align:middle line:84% AUDIENCE: How far would you split into three rules 00:52:08.820 --> 00:52:09.320 align:middle line:90% here, right? 00:52:09.320 --> 00:52:09.770 align:middle line:90% NORVIN RICHARDS: Yeah. 00:52:09.770 --> 00:52:11.273 align:middle line:84% AUDIENCE: First of all going around the two 00:52:11.273 --> 00:52:13.856 align:middle line:84% stressed syllables in a row, so I'm not sure on the second one 00:52:13.856 --> 00:52:14.690 align:middle line:90% now. 00:52:14.690 --> 00:52:15.240 align:middle line:90% NORVIN RICHARDS: Yeah. 00:52:15.240 --> 00:52:17.115 align:middle line:84% AUDIENCE: And don't stress the last syllable, 00:52:17.115 --> 00:52:18.710 align:middle line:84% like going to the three dot three. 00:52:18.710 --> 00:52:21.710 align:middle line:84% So I dictate, my mind goes very close somewhere. 00:52:21.710 --> 00:52:24.680 align:middle line:84% NORVIN RICHARDS: Yes, yes, yes, I see what you mean. 00:52:24.680 --> 00:52:27.680 align:middle line:84% So we could ask ourselves, what would happen if you had-- 00:52:27.680 --> 00:52:29.420 align:middle line:84% and this is either a virtue or a flaw 00:52:29.420 --> 00:52:31.010 align:middle line:84% of this way of talking about it, what 00:52:31.010 --> 00:52:33.560 align:middle line:84% would happen if the second thing there, don't stress 00:52:33.560 --> 00:52:36.890 align:middle line:84% the last syllable, were the first constraint, 00:52:36.890 --> 00:52:38.383 align:middle line:84% or if it were the third constraint, 00:52:38.383 --> 00:52:39.800 align:middle line:84% or if it were the last constraint. 00:52:39.800 --> 00:52:41.967 align:middle line:84% We get to ask ourselves, do we get different systems 00:52:41.967 --> 00:52:43.708 align:middle line:84% if we mess with the constraints that way? 00:52:43.708 --> 00:52:45.500 align:middle line:84% And this is indeed the kind of thing people 00:52:45.500 --> 00:52:48.480 align:middle line:84% do when they're doing this kind of work, you know. 00:52:48.480 --> 00:52:52.930 align:middle line:90% 00:52:52.930 --> 00:52:54.860 align:middle line:90% Here's another stress system. 00:52:54.860 --> 00:52:59.650 align:middle line:84% This is Passamaquoddy, in which the rules 00:52:59.650 --> 00:53:04.870 align:middle line:84% go, stress the first syllable, don't stress the last syllable. 00:53:04.870 --> 00:53:07.210 align:middle line:84% Those are the most important things. 00:53:07.210 --> 00:53:10.690 align:middle line:84% You can see that those are true in all of these examples. 00:53:10.690 --> 00:53:16.270 align:middle line:84% And don't allow two unstressed syllables in a row. 00:53:16.270 --> 00:53:18.310 align:middle line:84% And then down at the bottom of the list 00:53:18.310 --> 00:53:20.350 align:middle line:84% is don't allow two stressed syllables in a row. 00:53:20.350 --> 00:53:23.140 align:middle line:84% And it's at the bottom because, well, Passamaquoddy 00:53:23.140 --> 00:53:27.813 align:middle line:84% has two stressed syllables in a row all the time. 00:53:27.813 --> 00:53:29.980 align:middle line:84% Impressionistically what's going on in Passamaquoddy 00:53:29.980 --> 00:53:31.390 align:middle line:84% is you stress the first syllable, 00:53:31.390 --> 00:53:33.700 align:middle line:84% and you stress every other syllable counting backwards 00:53:33.700 --> 00:53:35.900 align:middle line:90% from the end of the word. 00:53:35.900 --> 00:53:41.690 align:middle line:84% So "dirt" is "tupqan," and then "in the dirt" is "tupqanok." 00:53:41.690 --> 00:53:43.810 align:middle line:84% So you get stress on the first two syllables, 00:53:43.810 --> 00:53:48.340 align:middle line:84% more stress actually on the second one than on the first. 00:53:48.340 --> 00:53:50.547 align:middle line:84% I wish that this were not the Passamaquoddy stress 00:53:50.547 --> 00:53:52.630 align:middle line:84% system, because I'm trying to learn Passamaquoddy. 00:53:52.630 --> 00:53:54.880 align:middle line:84% And it means that if you're trying 00:53:54.880 --> 00:53:57.970 align:middle line:84% to say something like "Let's walk around on top," 00:53:57.970 --> 00:53:59.890 align:middle line:84% you have to ask yourself, OK, is there 00:53:59.890 --> 00:54:02.820 align:middle line:84% an odd number of syllables in that word or an even number, 00:54:02.820 --> 00:54:04.570 align:middle line:84% because that's how you know whether to say 00:54:04.570 --> 00:54:08.560 align:middle line:84% "tehsahqapasultine," with stress, 00:54:08.560 --> 00:54:11.800 align:middle line:84% main stress on the next-to-last syllable, with two stressed 00:54:11.800 --> 00:54:15.860 align:middle line:84% syllables at the beginning, or not. 00:54:15.860 --> 00:54:18.810 align:middle line:90% Yeah, so it's a pain. 00:54:18.810 --> 00:54:27.021 align:middle line:84% Yeah, or here's Tauya, where, again, we could decompose this. 00:54:27.021 --> 00:54:29.910 align:middle line:84% This is not unlike Passamaquoddy, 00:54:29.910 --> 00:54:32.610 align:middle line:84% except you're stressing the first syllable. 00:54:32.610 --> 00:54:36.060 align:middle line:84% And you are stressing the last syllable. 00:54:36.060 --> 00:54:38.520 align:middle line:84% And then you are stressing every other syllable counting 00:54:38.520 --> 00:54:40.350 align:middle line:90% backwards from the end. 00:54:40.350 --> 00:54:43.840 align:middle line:84% So you end up with bi-syllables having two stresses, 00:54:43.840 --> 00:54:45.840 align:middle line:84% unlike in Passamaquoddy, where there's only one. 00:54:45.840 --> 00:54:48.695 align:middle line:90% 00:54:48.695 --> 00:54:50.070 align:middle line:84% That's one kind of stress system, 00:54:50.070 --> 00:54:52.440 align:middle line:84% these kinds of stress systems where you have repeated 00:54:52.440 --> 00:54:53.850 align:middle line:90% stresses. 00:54:53.850 --> 00:54:56.085 align:middle line:84% There's another kind of stress system 00:54:56.085 --> 00:54:58.710 align:middle line:84% that's going to have to include constraints that make reference 00:54:58.710 --> 00:55:00.100 align:middle line:90% to particular morphemes. 00:55:00.100 --> 00:55:02.610 align:middle line:84% So I want to give you an example of that. 00:55:02.610 --> 00:55:05.280 align:middle line:84% The classic example that I know of is from Russian. 00:55:05.280 --> 00:55:09.060 align:middle line:84% So I'm going to show you some Russian data. 00:55:09.060 --> 00:55:10.560 align:middle line:84% English has examples like this, too. 00:55:10.560 --> 00:55:12.143 align:middle line:84% We've talked about it before, the fact 00:55:12.143 --> 00:55:14.700 align:middle line:84% that "electric" has stress on its second syllable, 00:55:14.700 --> 00:55:16.970 align:middle line:84% but that when you add the suffix "ity," 00:55:16.970 --> 00:55:18.750 align:middle line:84% you do various interesting things. 00:55:18.750 --> 00:55:21.270 align:middle line:84% You change the "k" sound at the end of electric 00:55:21.270 --> 00:55:23.140 align:middle line:84% to an "s," so you get "electricity." 00:55:23.140 --> 00:55:27.870 align:middle line:84% But also you shift the stress, so it's not "electricity," 00:55:27.870 --> 00:55:32.400 align:middle line:84% it's "electricity" with stress on what was an "ik" and is now 00:55:32.400 --> 00:55:34.380 align:middle line:90% an "is." 00:55:34.380 --> 00:55:38.310 align:middle line:84% For Russian, the basic stress rule goes something like this. 00:55:38.310 --> 00:55:41.100 align:middle line:84% Russian morphemes get to indicate 00:55:41.100 --> 00:55:43.770 align:middle line:84% whether they want stress or not, and where 00:55:43.770 --> 00:55:45.330 align:middle line:90% they want the stress to be. 00:55:45.330 --> 00:55:48.803 align:middle line:84% And the basic rule goes, stress the first lexically-accented 00:55:48.803 --> 00:55:50.970 align:middle line:84% syllable, that is, the first syllable that is marked 00:55:50.970 --> 00:55:52.680 align:middle line:90% as saying, hey, I need stress. 00:55:52.680 --> 00:55:55.200 align:middle line:84% If there aren't any lexically-accented syllables, 00:55:55.200 --> 00:55:57.660 align:middle line:84% then you stress the first syllable of the word. 00:55:57.660 --> 00:56:00.210 align:middle line:84% Russian stress is quite intimidating 00:56:00.210 --> 00:56:01.920 align:middle line:84% if you're trying to learn Russian. 00:56:01.920 --> 00:56:04.145 align:middle line:84% So if you're interested in trying to learn Russian, 00:56:04.145 --> 00:56:06.270 align:middle line:84% please try to remember this rule that will save you 00:56:06.270 --> 00:56:07.990 align:middle line:90% a lot of heartache. 00:56:07.990 --> 00:56:10.120 align:middle line:90% So here are some examples. 00:56:10.120 --> 00:56:14.257 align:middle line:84% The word for town, "gorod," is an unaccented word. 00:56:14.257 --> 00:56:16.590 align:middle line:84% And that means that it has stress on the first syllable, 00:56:16.590 --> 00:56:19.320 align:middle line:84% because the rule is stress the first syllable if there aren't 00:56:19.320 --> 00:56:21.150 align:middle line:84% any lexically-accented syllables. 00:56:21.150 --> 00:56:23.280 align:middle line:84% And the dative singular suffix, "-u," 00:56:23.280 --> 00:56:25.650 align:middle line:84% that's showing up for this class of nouns, 00:56:25.650 --> 00:56:28.140 align:middle line:84% is also lexically unaccented, so it doesn't do anything 00:56:28.140 --> 00:56:28.870 align:middle line:90% to the stress. 00:56:28.870 --> 00:56:31.410 align:middle line:84% So "gorod" is the word for town, and "gorodu" 00:56:31.410 --> 00:56:32.430 align:middle line:90% is the dative singular. 00:56:32.430 --> 00:56:34.890 align:middle line:84% And again, don't worry about what dative singular means. 00:56:34.890 --> 00:56:35.922 align:middle line:90% It's a thing. 00:56:35.922 --> 00:56:37.380 align:middle line:84% It's a marker you can put on nouns. 00:56:37.380 --> 00:56:40.540 align:middle line:90% 00:56:40.540 --> 00:56:44.440 align:middle line:84% But the dative plural suffix, putting it in a different form, 00:56:44.440 --> 00:56:47.350 align:middle line:90% is a lexically accented stress. 00:56:47.350 --> 00:56:48.820 align:middle line:90% It does require stress. 00:56:48.820 --> 00:56:51.460 align:middle line:84% And so stress shifts to that particular suffix. 00:56:51.460 --> 00:56:53.410 align:middle line:84% So you get initial stress because there's 00:56:53.410 --> 00:56:55.810 align:middle line:84% no lexical accent in gorod and gorodu, 00:56:55.810 --> 00:56:57.550 align:middle line:84% but if you add the data plural suffix, 00:56:57.550 --> 00:57:00.880 align:middle line:84% you get "gorodam," with the stress shifting to the end, 00:57:00.880 --> 00:57:03.850 align:middle line:84% because this is a suffix that demands stress. 00:57:03.850 --> 00:57:05.470 align:middle line:84% There are also nouns that want stress 00:57:05.470 --> 00:57:06.850 align:middle line:90% on a particular syllable. 00:57:06.850 --> 00:57:10.780 align:middle line:84% So the word for nut, which is "orex," wants-- 00:57:10.780 --> 00:57:12.970 align:middle line:84% has lexical accent on its second syllable. 00:57:12.970 --> 00:57:16.200 align:middle line:84% And that means that, no matter what you add to that noun, 00:57:16.200 --> 00:57:17.980 align:middle line:84% the stress is not going to shift. 00:57:17.980 --> 00:57:20.440 align:middle line:90% So the word for nut is "orex." 00:57:20.440 --> 00:57:23.200 align:middle line:84% And if it's dative singular, if it's dative plural, 00:57:23.200 --> 00:57:25.300 align:middle line:84% you still get stress on that second syllable, 00:57:25.300 --> 00:57:28.450 align:middle line:84% because the basic rule is stress the first lexically-accented 00:57:28.450 --> 00:57:29.440 align:middle line:90% syllable. 00:57:29.440 --> 00:57:33.037 align:middle line:84% So in the last example, the dative plural of nut, 00:57:33.037 --> 00:57:34.870 align:middle line:84% the stress is still in that second syllable. 00:57:34.870 --> 00:57:37.420 align:middle line:84% It's "orexam," even though, as we've seen, "am" 00:57:37.420 --> 00:57:38.860 align:middle line:90% attracts stress to itself. 00:57:38.860 --> 00:57:40.360 align:middle line:84% It only attracts stress to itself 00:57:40.360 --> 00:57:43.528 align:middle line:84% from nouns that don't themselves have lexical accent, yeah. 00:57:43.528 --> 00:57:45.070 align:middle line:84% There, now you know more about that-- 00:57:45.070 --> 00:57:47.230 align:middle line:84% I tried years ago to try to learn Russian. 00:57:47.230 --> 00:57:49.450 align:middle line:84% And this was one of the things that made me give up. 00:57:49.450 --> 00:57:51.850 align:middle line:84% It was like the stress just doesn't make sense to me. 00:57:51.850 --> 00:57:53.990 align:middle line:90% I don't understand it. 00:57:53.990 --> 00:57:56.080 align:middle line:84% But now you know, so and I do too. 00:57:56.080 --> 00:57:58.780 align:middle line:84% I should go try to learn it again. 00:57:58.780 --> 00:58:00.355 align:middle line:84% So one way to think about this is 00:58:00.355 --> 00:58:04.330 align:middle line:84% you say a Russian morpheme can have a lexical accent. 00:58:04.330 --> 00:58:08.510 align:middle line:84% As part of its specification, it can have a little note saying, 00:58:08.510 --> 00:58:10.670 align:middle line:90% hey, I want stress to be here. 00:58:10.670 --> 00:58:12.910 align:middle line:84% And then we need constraints that say things like, 00:58:12.910 --> 00:58:15.700 align:middle line:84% there's only one stress in a word. 00:58:15.700 --> 00:58:17.800 align:middle line:84% You want to stress lexical accents, 00:58:17.800 --> 00:58:21.350 align:middle line:84% and you should put your stress as far to the left as possible. 00:58:21.350 --> 00:58:25.840 align:middle line:84% So in a word like town, there aren't any lexical accents. 00:58:25.840 --> 00:58:28.633 align:middle line:84% And so the only relevant constraints, I guess, 00:58:28.633 --> 00:58:29.800 align:middle line:90% are the first and the third. 00:58:29.800 --> 00:58:31.030 align:middle line:84% There should be one stress, and it should 00:58:31.030 --> 00:58:32.200 align:middle line:90% be as far left as possible. 00:58:32.200 --> 00:58:35.410 align:middle line:84% So it's at the beginning, "gorod." 00:58:35.410 --> 00:58:39.220 align:middle line:84% For "gorodam," the dative plural of town, 00:58:39.220 --> 00:58:40.780 align:middle line:90% there can only be one stress. 00:58:40.780 --> 00:58:42.530 align:middle line:84% You want it to be as far left as possible. 00:58:42.530 --> 00:58:45.340 align:middle line:84% But there's a lexical accent on dative plural suffix. 00:58:45.340 --> 00:58:47.590 align:middle line:90% And so you put it on "am." 00:58:47.590 --> 00:58:50.020 align:middle line:84% that's as far left as you can go, yeah. 00:58:50.020 --> 00:58:52.870 align:middle line:84% So the last constraint doesn't get to do very much. 00:58:52.870 --> 00:58:56.320 align:middle line:84% And then for nut, there's lexical accent 00:58:56.320 --> 00:58:58.600 align:middle line:84% on the "ex," on the second syllable of the noun. 00:58:58.600 --> 00:59:01.275 align:middle line:84% And because you're putting stress as far left as possible, 00:59:01.275 --> 00:59:02.650 align:middle line:84% that's where the stress is always 00:59:02.650 --> 00:59:04.540 align:middle line:84% going to go, no matter what suffixes you add. 00:59:04.540 --> 00:59:06.250 align:middle line:90% Yeah, that's the basic picture. 00:59:06.250 --> 00:59:09.080 align:middle line:90% 00:59:09.080 --> 00:59:13.670 align:middle line:84% OK, we will do a little bit of Kashmiri and then we'll stop, 00:59:13.670 --> 00:59:15.410 align:middle line:90% I think, for today. 00:59:15.410 --> 00:59:18.050 align:middle line:84% This actually came up when we were talking about cucumbers. 00:59:18.050 --> 00:59:20.210 align:middle line:84% There are languages that have stress systems 00:59:20.210 --> 00:59:22.470 align:middle line:84% that care about what's called syllable weight. 00:59:22.470 --> 00:59:25.910 align:middle line:84% That is, they care about whether vowels are long or short 00:59:25.910 --> 00:59:27.950 align:middle line:84% and also about how many consonants there 00:59:27.950 --> 00:59:30.650 align:middle line:90% are after vowels. 00:59:30.650 --> 00:59:33.270 align:middle line:84% And Kashmiri is a classic example of this. 00:59:33.270 --> 00:59:35.720 align:middle line:84% And I'll show you a little bit of this 00:59:35.720 --> 00:59:38.030 align:middle line:84% and we'll see whether we can find a good stopping 00:59:38.030 --> 00:59:41.420 align:middle line:90% place somewhere in the middle. 00:59:41.420 --> 00:59:42.680 align:middle line:90% Here are some Kashmiri words. 00:59:42.680 --> 00:59:45.097 align:middle line:84% There are lots of Kashmiri words that have initial stress. 00:59:45.097 --> 00:59:48.562 align:middle line:84% Anybody here speak Kashmiri, language of Kashmir? 00:59:48.562 --> 00:59:51.450 align:middle line:90% OK. 00:59:51.450 --> 00:59:53.312 align:middle line:84% So lots of words have initial stress. 00:59:53.312 --> 00:59:55.020 align:middle line:84% Basically, initial stress is what you get 00:59:55.020 --> 00:59:57.250 align:middle line:90% if nothing else is happening. 00:59:57.250 --> 00:59:58.920 align:middle line:84% But if there's a long vowel in the word, 00:59:58.920 --> 01:00:00.920 align:middle line:84% even if it's not the first syllable of the word, 01:00:00.920 --> 01:00:03.330 align:middle line:84% then the long vowel will get stress. 01:00:03.330 --> 01:00:07.653 align:middle line:84% So that verb, to finish, has a long vowel as its third vowel, 01:00:07.653 --> 01:00:09.570 align:middle line:84% and that's where the stress goes in that word. 01:00:09.570 --> 01:00:12.930 align:middle line:90% 01:00:12.930 --> 01:00:15.750 align:middle line:84% A long vowel will attract stress except when it's final. 01:00:15.750 --> 01:00:20.230 align:middle line:84% So if it's the last vowel in the word, it doesn't get stressed. 01:00:20.230 --> 01:00:22.382 align:middle line:84% So the word for book has two syllables. 01:00:22.382 --> 01:00:24.090 align:middle line:84% And the second syllable has a long vowel. 01:00:24.090 --> 01:00:27.075 align:middle line:84% But the first syllable gets the stress, because Kashmiri 01:00:27.075 --> 01:00:28.950 align:middle line:84% like several other languages we've looked at, 01:00:28.950 --> 01:00:32.050 align:middle line:84% doesn't like stressing final syllables. 01:00:32.050 --> 01:00:36.820 align:middle line:84% If there are two long vowels, then you stress the first one. 01:00:36.820 --> 01:00:38.755 align:middle line:84% And another kind of thing that attracts stress 01:00:38.755 --> 01:00:40.630 align:middle line:84% is a vowel that's followed by two consonants. 01:00:40.630 --> 01:00:44.320 align:middle line:84% So here's the word for "Friday," where the second vowel is 01:00:44.320 --> 01:00:46.038 align:middle line:90% followed by an "r" and a "v." 01:00:46.038 --> 01:00:47.830 align:middle line:84% It's the first example, I guess, that we've 01:00:47.830 --> 01:00:50.337 align:middle line:84% had of a vowel that's followed by two consonants. 01:00:50.337 --> 01:00:52.420 align:middle line:84% And a vowel that's followed by two consonants also 01:00:52.420 --> 01:00:54.160 align:middle line:90% preferentially gets stressed. 01:00:54.160 --> 01:00:57.080 align:middle line:90% It acts like a long vowel. 01:00:57.080 --> 01:00:59.080 align:middle line:84% If you have both a long vowel and a vowel that's 01:00:59.080 --> 01:01:01.900 align:middle line:84% followed by two consonants, then you prefer the long vowel. 01:01:01.900 --> 01:01:05.830 align:middle line:84% So you're getting that in here in door. 01:01:05.830 --> 01:01:08.185 align:middle line:84% In Friday, you're not getting stress on the long vowel. 01:01:08.185 --> 01:01:10.810 align:middle line:84% That's because the long vowel is the last syllable of the word, 01:01:10.810 --> 01:01:14.600 align:middle line:84% and the more important thing is don't stress the last syllable. 01:01:14.600 --> 01:01:17.380 align:middle line:84% So we can fruitfully think of Kashmiri stress 01:01:17.380 --> 01:01:18.820 align:middle line:84% as a bunch of ranked constraints, 01:01:18.820 --> 01:01:21.400 align:middle line:84% again, which say things like don't 01:01:21.400 --> 01:01:22.790 align:middle line:90% have more than one stress. 01:01:22.790 --> 01:01:24.640 align:middle line:90% Don't stress the final syllable. 01:01:24.640 --> 01:01:26.260 align:middle line:90% Stress your long vowels. 01:01:26.260 --> 01:01:28.870 align:middle line:84% Stress your vowels that have multiple consonants after them, 01:01:28.870 --> 01:01:30.820 align:middle line:84% and then put stress as far to the left 01:01:30.820 --> 01:01:32.770 align:middle line:90% in the word as possible. 01:01:32.770 --> 01:01:35.980 align:middle line:84% So stress long vowels, we got that in that word 01:01:35.980 --> 01:01:39.490 align:middle line:84% that had a long vowel as its third vowel. 01:01:39.490 --> 01:01:42.310 align:middle line:84% But we also saw that outranking stress 01:01:42.310 --> 01:01:44.007 align:middle line:84% long vowels is a constraint that says 01:01:44.007 --> 01:01:45.340 align:middle line:90% don't stress the final syllable. 01:01:45.340 --> 01:01:47.493 align:middle line:84% So in this word for "book," there's 01:01:47.493 --> 01:01:48.910 align:middle line:84% a long vowel in the final syllable 01:01:48.910 --> 01:01:50.743 align:middle line:84% but you don't stress it, because it's final. 01:01:50.743 --> 01:01:53.230 align:middle line:84% And avoiding final stress is more important than stressing 01:01:53.230 --> 01:01:55.540 align:middle line:90% long vowels. 01:01:55.540 --> 01:01:58.090 align:middle line:84% There's a rule that says don't have more than one stress. 01:01:58.090 --> 01:02:00.402 align:middle line:84% So you like stressing long vowels, 01:02:00.402 --> 01:02:02.860 align:middle line:84% but if you have a word that has multiple long vowels in it, 01:02:02.860 --> 01:02:03.940 align:middle line:90% you only get one stress. 01:02:03.940 --> 01:02:06.910 align:middle line:90% You don't get two, yeah? 01:02:06.910 --> 01:02:09.660 align:middle line:84% And the general rule is, put stress as far to the left 01:02:09.660 --> 01:02:10.660 align:middle line:90% in the word as possible. 01:02:10.660 --> 01:02:12.910 align:middle line:84% So when you have two long vowels you stress the first one. 01:02:12.910 --> 01:02:13.990 align:middle line:90% You can't stress them both. 01:02:13.990 --> 01:02:15.370 align:middle line:84% You also don't stress the second one. 01:02:15.370 --> 01:02:17.370 align:middle line:84% You stress the first one because you want stress 01:02:17.370 --> 01:02:19.120 align:middle line:84% to be as far to the left as possible. 01:02:19.120 --> 01:02:21.820 align:middle line:84% That's also why, when there are no long vowels, and no vowels 01:02:21.820 --> 01:02:25.270 align:middle line:84% with multiple consonants after them, you get initial stress. 01:02:25.270 --> 01:02:27.820 align:middle line:84% And then finally, stressed vowels with multiple consonants 01:02:27.820 --> 01:02:30.070 align:middle line:84% after them, that's got to be in there somewhere. 01:02:30.070 --> 01:02:32.500 align:middle line:84% And it's got to be below stressed long vowels, 01:02:32.500 --> 01:02:36.190 align:middle line:84% because when you have both a vowel with multiple consonants 01:02:36.190 --> 01:02:39.213 align:middle line:84% after it and a long vowel, you stress the long vowel 01:02:39.213 --> 01:02:41.380 align:middle line:84% and not the vowel with multiple consonants after it. 01:02:41.380 --> 01:02:43.900 align:middle line:84% So here's another place where it's 01:02:43.900 --> 01:02:49.750 align:middle line:84% possible to describe Kashmiri stress in several paragraphs 01:02:49.750 --> 01:02:52.930 align:middle line:84% of prose, in which you say things 01:02:52.930 --> 01:02:56.200 align:middle line:84% like, well, Kashmiri stress, it's on the first syllable 01:02:56.200 --> 01:02:59.080 align:middle line:84% unless there are long vowels or vowels with multiple consonants 01:02:59.080 --> 01:03:01.300 align:middle line:84% after them, except when those are final 01:03:01.300 --> 01:03:04.525 align:middle line:84% or on a Tuesday in a month beginning with "m." 01:03:04.525 --> 01:03:08.080 align:middle line:84% It would be quite long and complicated. 01:03:08.080 --> 01:03:10.240 align:middle line:84% If we describe Kashmiri stress this way in terms 01:03:10.240 --> 01:03:12.430 align:middle line:84% of ranked constraints, the constraints themselves 01:03:12.430 --> 01:03:13.640 align:middle line:90% can all be pretty simple. 01:03:13.640 --> 01:03:15.640 align:middle line:84% They can be things like stress your long vowels. 01:03:15.640 --> 01:03:17.740 align:middle line:84% That's a pretty natural thing to do. 01:03:17.740 --> 01:03:19.450 align:middle line:90% Or avoid final stress. 01:03:19.450 --> 01:03:22.670 align:middle line:84% Or put your stress as far to the left in the word as possible. 01:03:22.670 --> 01:03:24.910 align:middle line:84% And all of the action is figuring out 01:03:24.910 --> 01:03:26.415 align:middle line:84% what order are these constraints in, 01:03:26.415 --> 01:03:27.790 align:middle line:84% which are the most important ones 01:03:27.790 --> 01:03:29.020 align:middle line:90% and which are less important. 01:03:29.020 --> 01:03:30.400 align:middle line:84% And by ranking them that way, you 01:03:30.400 --> 01:03:32.943 align:middle line:84% end up with this quite complicated array 01:03:32.943 --> 01:03:35.110 align:middle line:84% of data, which we've now gone through quite quickly. 01:03:35.110 --> 01:03:37.090 align:middle line:90% But you'll have the slides. 01:03:37.090 --> 01:03:40.510 align:middle line:84% So you'll have a chance to go through this in your own time. 01:03:40.510 --> 01:03:42.130 align:middle line:84% Are the questions about Kashmiri? 01:03:42.130 --> 01:03:43.780 align:middle line:84% Since we have a few more seconds here, 01:03:43.780 --> 01:03:45.520 align:middle line:84% we can talk about Kashmiri in more depth 01:03:45.520 --> 01:03:48.332 align:middle line:84% if people would like, or anything else. 01:03:48.332 --> 01:03:50.290 align:middle line:84% As I said at the beginning, I could hear people 01:03:50.290 --> 01:03:51.190 align:middle line:90% talking about the problem set. 01:03:51.190 --> 01:03:52.570 align:middle line:84% If people have questions about the problem set, 01:03:52.570 --> 01:03:53.528 align:middle line:90% we can talk about them. 01:03:53.528 --> 01:03:54.195 align:middle line:90% Yeah. 01:03:54.195 --> 01:03:57.550 align:middle line:84% AUDIENCE: The chapter felt like-- so is, like, 01:03:57.550 --> 01:03:58.270 align:middle line:90% for like English. 01:03:58.270 --> 01:03:59.170 align:middle line:90% NORVIN RICHARDS: Yeah. 01:03:59.170 --> 01:04:00.670 align:middle line:84% AUDIENCE: Is there like a system like this? 01:04:00.670 --> 01:04:01.090 align:middle line:90% NORVIN RICHARDS: Oh. 01:04:01.090 --> 01:04:01.780 align:middle line:84% AUDIENCE: Or is it just kind of like-- 01:04:01.780 --> 01:04:02.500 align:middle line:84% NORVIN RICHARDS: So, yeah, here I 01:04:02.500 --> 01:04:03.850 align:middle line:90% am talking about stress systems. 01:04:03.850 --> 01:04:05.538 align:middle line:84% And I've talked to you about Kashmiri 01:04:05.538 --> 01:04:07.330 align:middle line:84% and I've talked to you about Passamaquoddy. 01:04:07.330 --> 01:04:09.497 align:middle line:84% I've talked to you about various languages, Russian, 01:04:09.497 --> 01:04:13.270 align:middle line:84% and I've strenuously avoided talking to you about English. 01:04:13.270 --> 01:04:16.820 align:middle line:84% English stress is quite complicated and messy. 01:04:16.820 --> 01:04:18.980 align:middle line:90% There are books about it. 01:04:18.980 --> 01:04:21.310 align:middle line:84% It's partly, like so much about English, 01:04:21.310 --> 01:04:26.470 align:middle line:84% the product of several different languages with different stress 01:04:26.470 --> 01:04:29.600 align:middle line:84% systems interacting with each other. 01:04:29.600 --> 01:04:33.280 align:middle line:84% So there was a Germanic stress system 01:04:33.280 --> 01:04:36.760 align:middle line:84% that English inherited from its Germanic substrate. 01:04:36.760 --> 01:04:40.270 align:middle line:84% And then the French invaded in 1066 01:04:40.270 --> 01:04:42.110 align:middle line:84% bringing their own stress system. 01:04:42.110 --> 01:04:45.130 align:middle line:84% And one of the results of that is 01:04:45.130 --> 01:04:47.200 align:middle line:84% that English stress is the kind of thing 01:04:47.200 --> 01:04:48.462 align:middle line:90% you can write books about. 01:04:48.462 --> 01:04:50.170 align:middle line:84% So there's a reason I haven't tried to do 01:04:50.170 --> 01:04:51.520 align:middle line:90% English stress on the slide. 01:04:51.520 --> 01:04:54.790 align:middle line:84% You think Kashmiri is complicated, yeah. 01:04:54.790 --> 01:04:59.050 align:middle line:90% It's, yeah, it's complicated. 01:04:59.050 --> 01:05:00.640 align:middle line:84% It's complicated in interesting ways. 01:05:00.640 --> 01:05:02.420 align:middle line:84% We have various places, for example, 01:05:02.420 --> 01:05:05.110 align:middle line:84% where the difference between a verb and a noun 01:05:05.110 --> 01:05:06.700 align:middle line:90% has to do with stress. 01:05:06.700 --> 01:05:12.220 align:middle line:84% So we have verbs like "perMIT" and nouns like "PERmit," right? 01:05:12.220 --> 01:05:15.100 align:middle line:90% Get a permit to do something. 01:05:15.100 --> 01:05:17.850 align:middle line:84% And there are a number of pairs like that, 01:05:17.850 --> 01:05:19.600 align:middle line:84% which is one of the kinds of things that's 01:05:19.600 --> 01:05:20.725 align:middle line:90% interesting to think about. 01:05:20.725 --> 01:05:22.540 align:middle line:84% That's not something that we inherited. 01:05:22.540 --> 01:05:27.010 align:middle line:84% And so when people are trying to figure out how-- 01:05:27.010 --> 01:05:29.680 align:middle line:84% we talked earlier about the possibility 01:05:29.680 --> 01:05:32.170 align:middle line:84% that nouns and verbs could have different phonologies, 01:05:32.170 --> 01:05:34.300 align:middle line:84% and there are times when something like that 01:05:34.300 --> 01:05:37.330 align:middle line:84% seems to be happening, like in that kind of example. 01:05:37.330 --> 01:05:39.988 align:middle line:84% So when you write your book on English stress, that's 01:05:39.988 --> 01:05:41.530 align:middle line:84% one of the chapters you have to write 01:05:41.530 --> 01:05:43.570 align:middle line:84% is the difference between permit and permit. 01:05:43.570 --> 01:05:44.530 align:middle line:90% Yeah. 01:05:44.530 --> 01:05:45.298 align:middle line:90% Yeah. 01:05:45.298 --> 01:05:47.173 align:middle line:84% AUDIENCE: You said, is there like some method 01:05:47.173 --> 01:05:49.360 align:middle line:84% we can use to define the like all the rules 01:05:49.360 --> 01:05:50.110 align:middle line:90% that were in this? 01:05:50.110 --> 01:05:52.770 align:middle line:84% Like adjusting rules to see what the basis is. 01:05:52.770 --> 01:05:54.960 align:middle line:84% NORVIN RICHARDS: So what you want to do 01:05:54.960 --> 01:05:57.270 align:middle line:84% is look systematically at each vowel. 01:05:57.270 --> 01:05:59.520 align:middle line:84% Here's what I would do in your shoes. 01:05:59.520 --> 01:06:01.290 align:middle line:84% Look systematically at each vowel 01:06:01.290 --> 01:06:05.010 align:middle line:84% and make yourself notes for each vowel about the environments 01:06:05.010 --> 01:06:06.480 align:middle line:90% that they show up in. 01:06:06.480 --> 01:06:11.040 align:middle line:84% So, say you start with the vowel "a." 01:06:11.040 --> 01:06:14.952 align:middle line:84% Make yourself notes about all of the things that are around "a." 01:06:14.952 --> 01:06:16.410 align:middle line:84% Maybe you could start with the hope 01:06:16.410 --> 01:06:18.977 align:middle line:84% that the relevant things are the consonants on either side 01:06:18.977 --> 01:06:21.060 align:middle line:84% of it, or whether it's at the beginning or the end 01:06:21.060 --> 01:06:23.435 align:middle line:84% or the first syllable or the second syllable or whatever. 01:06:23.435 --> 01:06:26.730 align:middle line:84% Make yourself notes for each word that 01:06:26.730 --> 01:06:29.228 align:middle line:84% has an "i" in it, where the "i" is showing up. 01:06:29.228 --> 01:06:31.770 align:middle line:84% And then look at your notes and hope that something jumps out 01:06:31.770 --> 01:06:32.700 align:middle line:90% at you. 01:06:32.700 --> 01:06:34.782 align:middle line:84% That's the only thing I can think of to suggest. 01:06:34.782 --> 01:06:36.490 align:middle line:84% Sorry, that's probably what you're doing. 01:06:36.490 --> 01:06:39.256 align:middle line:84% But if that's what you're doing, you're doing well. 01:06:39.256 --> 01:06:42.810 align:middle line:90% Are there other questions? 01:06:42.810 --> 01:06:45.630 align:middle line:84% OK, all right, so I'll see you on Thursday 01:06:45.630 --> 01:06:48.230 align:middle line:84% when we will start talking about syntax. 01:06:48.230 --> 01:06:57.000 align:middle line:90%