WEBVTT
Kind: captions
Language: en

1
00:00:00.089 --> 00:00:03.789
English spelling is a mess.
O-u-g-h becomes different sounds in

2
00:00:03.789 --> 00:00:07.610
enough, cough, plough, hiccough, although,
thought, and thoroughly.

3
00:00:07.610 --> 00:00:10.800
Part of the reason for that is that English
has so many loan words

4
00:00:10.800 --> 00:00:15.080
whose pronunciations got either maintained
or approximated.

5
00:00:15.080 --> 00:00:18.630
Some were written down in the way monks would
spell that sound

6
00:00:18.630 --> 00:00:21.060
whenever that word happened
to enter the language.

7
00:00:21.060 --> 00:00:23.420
Plus, French scribes from the Norman Invasion

8
00:00:23.420 --> 00:00:27.539
respelled English words to match
French spelling rules. Plus,

9
00:00:27.539 --> 00:00:31.820
English spelling became standardized with
the implementation of the printing press in

10
00:00:31.820 --> 00:00:33.570
the 16th century,

11
00:00:33.570 --> 00:00:38.320
which was the start of Early Modern English,
but before the Great Vowel Shift.

12
00:00:38.320 --> 00:00:42.000
Which was… well, a lot of our vowels shifted.

13
00:00:42.000 --> 00:00:43.550
Between the confusing spelling of English,

14
00:00:43.550 --> 00:00:47.790
and the fact other European languages use
different letters for different sounds,

15
00:00:47.790 --> 00:00:50.250
we needed something for linguists to us

16
00:00:50.250 --> 00:00:54.840
to explain the pronunciation of words without
having to talk face-to-face.

17
00:00:54.840 --> 00:00:59.809
So in the late 19th century, the International
Phonetics Association worked together

18
00:00:59.809 --> 00:01:04.830
to create the beginnings of the International
Phonetic Alphabet, or IPA.

19
00:01:04.830 --> 00:01:06.560
A century later, and after a few revisions,

20
00:01:06.560 --> 00:01:09.029
it’s now used around the world by linguists,

21
00:01:09.029 --> 00:01:13.880
but also phonologists, speech pathologists,
singers, and loads more people.

22
00:01:13.880 --> 00:01:18.760
The goal of the IPA is to have a symbol for
every sound in every spoken language in the world.

23
00:01:18.760 --> 00:01:20.779
Now, it does have very Eurocentric roots.

24
00:01:20.779 --> 00:01:23.429
The alphabet itself is based off
the Roman  alphabet,

25
00:01:23.429 --> 00:01:25.869
with some Greek letters thrown in.

26
00:01:25.869 --> 00:01:28.999
So the closer your language is
to European ones,

27
00:01:28.999 --> 00:01:32.079
the more the transcriptions look like regular
Latin text.

28
00:01:32.079 --> 00:01:36.180
It was designed to be easy to type without
needing many custom letter plates

29
00:01:36.180 --> 00:01:37.900
for printing presses or typewriters,

30
00:01:37.900 --> 00:01:41.619
so it also has some rotated letters
and partial symbols.

31
00:01:41.619 --> 00:01:45.310
And the designers didn't want to have to use
a lot of diacritics;

32
00:01:45.310 --> 00:01:49.329
they wanted one symbol per sound,
and one sound per symbol.

33
00:01:49.329 --> 00:01:55.299
And… yeah. They did it. They put every sound
that humans use in speech on a chart.

34
00:01:55.299 --> 00:01:57.639
When you make a sound,
you push air out of your lungs

35
00:01:57.639 --> 00:02:00.880
and then you cause some sort of constriction
in your mouth or throat

36
00:02:00.880 --> 00:02:02.440
to create that particular sound.

37
00:02:02.440 --> 00:02:06.900
Two things that are really important are where
in your mouth and throat that constriction is:

38
00:02:06.909 --> 00:02:09.289
that’s called the place of articulation,

39
00:02:09.289 --> 00:02:11.569
and how much constriction you're making:

40
00:02:11.569 --> 00:02:13.720
that’s the manner of articulation.

41
00:02:13.720 --> 00:02:17.580
The chart is organised
based on those two factors.

42
00:02:17.580 --> 00:02:18.760
There’s also a chart for vowels,

43
00:02:18.769 --> 00:02:21.239
a chart of diacritics, and a chart of extensions,

44
00:02:21.239 --> 00:02:25.269
but we're going to focus on the main consonant
chart for this video.

45
00:02:25.269 --> 00:02:27.540
Because there’s variation from
dialect to dialect,

46
00:02:27.540 --> 00:02:29.670
and language to language, and person to person,

47
00:02:29.670 --> 00:02:31.820
the little nuances of each of those sounds

48
00:02:31.820 --> 00:02:35.400
can be even further
"narrowly transcribed" using diacritics.

49
00:02:35.400 --> 00:02:40.170
But these are, with a few exceptions,
all the consonants that humans use.

50
00:02:40.170 --> 00:02:43.109
Looking at the chart, you can see some
familiar symbols.

51
00:02:43.109 --> 00:02:46.409
These are a great guide to help you
find your way around.

52
00:02:46.409 --> 00:02:49.189
We can start at the top with the plosives:

53
00:02:49.189 --> 00:02:52.530
those are the sounds that you make
by stopping air flow somewhere,

54
00:02:52.530 --> 00:02:55.830
and then releasing it suddenly, like
a mini explosion.

55
00:02:55.830 --> 00:02:57.900
If you say some of the letters you
recognise off the top row,

56
00:02:57.900 --> 00:02:59.959
as you move left to right,

57
00:02:59.959 --> 00:03:04.190
you can feel those letters moving backwards
in your mouth: p, t, k.

58
00:03:04.190 --> 00:03:08.709
At the end of that row is the glottal stop,
which is closing off the airflow in your throat.

59
00:03:08.709 --> 00:03:11.530
It happens in English when we say "uh-oh".

60
00:03:11.530 --> 00:03:16.360
Depending on your accent, it also occurs in
"mountain", "Hawai'i", or "butter".

61
00:03:16.360 --> 00:03:21.310
The nasal line, next down, is all the sounds
that you make through your nose, m, n, ng.

62
00:03:21.310 --> 00:03:24.480
Then you have trills, which we don't have natively
in most dialects of English,

63
00:03:24.480 --> 00:03:30.569
and which I’m therefore not great at pronouncing,
other than a really over-the-top [rolled r].

64
00:03:30.569 --> 00:03:33.510
Then we have taps and flaps, which are like
those long trills,

65
00:03:33.510 --> 00:03:36.370
but a tap is the one-time equivalent.

66
00:03:36.370 --> 00:03:40.310
Next are the fricatives, where two parts touch.
So, s and z,

67
00:03:40.310 --> 00:03:43.110
where the two parts are the tip of your tongue
and  the alveolar ridge,

68
00:03:43.110 --> 00:03:46.180
which is the hard bit of the roof of your
mouth, just behind your top teeth.

69
00:03:46.180 --> 00:03:48.379
Informally, that is  the “pizza ridge”,

70
00:03:48.379 --> 00:03:52.049
because it’s the bit you burn when you bite
into pizza that’s too hot.

71
00:03:52.049 --> 00:03:54.510
And then finally, there are the approximants,

72
00:03:54.510 --> 00:03:56.909
where two parts don’t quite meet.

73
00:03:56.909 --> 00:03:58.779
Lateral means that air is
going round the sides,

74
00:03:58.779 --> 00:04:00.660
like in l or the Welsh ll.

75
00:04:00.660 --> 00:04:03.930
But: not every space in the table has a symbol.

76
00:04:03.930 --> 00:04:06.480
Some of those empty spaces have white backgrounds,

77
00:04:06.480 --> 00:04:09.260
and some of them are
completely grayed out.

78
00:04:09.260 --> 00:04:13.609
The ones with white backgrounds are possible
for humans to make,

79
00:04:13.609 --> 00:04:16.000
but they've not been given a symbol,
because linguists have never found them

80
00:04:16.000 --> 00:04:20.139
actually being used as sounds in
any of the world's languages.

81
00:04:20.139 --> 00:04:22.919
Sometimes they are found to exist by
some researcher somewhere,

82
00:04:22.919 --> 00:04:27.080
and if that happens then the International
Phonetics Association will add a symbol.

83
00:04:27.080 --> 00:04:30.590
The last time that happened was in 2005.

84
00:04:30.590 --> 00:04:33.910
But then there are shaded areas,
the shadow lands,

85
00:04:33.910 --> 00:04:37.440
the sounds that are theorized to be impossible
to produce.

86
00:04:37.440 --> 00:04:40.960
Imagine trying to produce a sound that's in
the place of your mouth

87
00:04:40.960 --> 00:04:43.260
where you make a k or a g,

88
00:04:43.260 --> 00:04:45.990
but using the same method that
you use to go [rolled r].

89
00:04:45.990 --> 00:04:49.340
You need some loose fold of tissue to make
that rapid trilling,

90
00:04:49.340 --> 00:04:52.580
and there's just no flesh loose enough to
bounce around

91
00:04:52.580 --> 00:04:55.640
right at the back of the roof of your mouth,
at what’s called the velum.

92
00:04:55.640 --> 00:04:59.280
You could, I guess, in theory,
curve your tongue back there,

93
00:04:59.280 --> 00:05:01.830
but only by holding it too tense
to actually make a  trill.

94
00:05:01.830 --> 00:05:06.290
So a velar trill is judged impossible.
By all means, try it.

95
00:05:06.290 --> 00:05:09.789
In the same way, if you try and cross the
place of your mouth and throat

96
00:05:09.789 --> 00:05:15.690
that you use to say "ha" or "uh-oh" with
the method you use to make the l sound,

97
00:05:15.690 --> 00:05:17.710
you're going to run into problems.

98
00:05:17.710 --> 00:05:23.210
For "la" or "el" you let air escape from your
mouth around the sides of your tongue: lateral.

99
00:05:23.210 --> 00:05:28.640
But for "ha" you're constricting the air down
in your throat at the vocal folds or glottis,

100
00:05:28.640 --> 00:05:33.449
and you can't get your tongue down your throat
to create that centre blockage.

101
00:05:33.449 --> 00:05:36.740
I mean, your tongue has that flexibility
in your mouth, 'l',

102
00:05:36.740 --> 00:05:40.759
but you can’t close your airway like you’re
making a figure 8 out of a straw,

103
00:05:40.759 --> 00:05:43.070
and keep the air going round the sides.

104
00:05:43.070 --> 00:05:46.580
Unless you’re choking on food in exactly
the wrong way,

105
00:05:46.580 --> 00:05:48.220
and if you are, that's not really speaking.

106
00:05:48.220 --> 00:05:51.180
A glottal lateral is judged impossible.

107
00:05:51.180 --> 00:05:53.830
All those shaded boxes have the same sorts
of limitations,

108
00:05:53.830 --> 00:05:58.340
where it's not just that they would be difficult
to pronounce, they are impossible.

109
00:05:58.340 --> 00:06:01.140
Those gray boxes are forbidden
from human speech,

110
00:06:01.150 --> 00:06:05.659
not by choice of some High Council or the
International Phonetic Association,

111
00:06:05.660 --> 00:06:09.420
but just because of the
limitations of being human.

112
00:06:10.930 --> 00:06:13.960
Sometimes we do write things that we can't
pronounce, like these:

113
00:06:13.960 --> 00:06:15.620
you can't hold the letter p for that long,

114
00:06:15.620 --> 00:06:18.560
or this where all those letter Es
are silent!

115
00:06:18.560 --> 00:06:20.990
For more on how writing is
different to speaking,

116
00:06:20.990 --> 00:06:24.320
my co-author Gretchen McCulloch has a new book called
Because Internet,

117
00:06:24.320 --> 00:06:26.780
it's available now,
and there are links in the description.