Foudations of Phonology

Download as pdf or txt
Download as pdf or txt
You are on page 1of 66

ELTE BTK DELG

Foundations of Phonology
Lecture notes Handouts Additional information

2013

These lecture notes are based on the lecture Foundations of Phonology presented by
Dr. Pter Szigetvri, associate professor, head of DELG; however EVERY mistake,
inaccuracy or any other kind of fault is the responsibility of the author (who is not
mentioned of course :D).
Consonants1
I Branches of phonetics: (some aspects of phonetics)
1 Physical properties of sounds: the study of the physical properties of speech sounds
(a.k.a. phonetics) has three branches:
a) articulatory ("dealing with how to utter"): what to put where and how to get the
given sound
b) acoustic ("dealing with how to be heard"): the properties of the sound waves of
the given sound
c) auditory ("dealing with how to hear"): how the brain processes the acoustic
signals the ear receives
2 Articulatory properties of consonants: (in general)
a) place of articulation (POA): the location within the vocal tract 2where the most
significant constriction3 occurs
b) manner of articulation: the type of constriction that occurs
c) phonation type: the configuration of
the larynx (the glottis) (gge, ggef)
d) airstream mechanism: the initiator4
and the direction of the airstream
3 Place of articulation:
a) articulators5:
passive:
( 2) lips/labial (ajakhang)
(3) teeth/dental (foghang)
(4) upper alveolar
ridge6/alveolar
(5) /post-alveolar
(6) /pre-palatal
(7) palate, hard palate7/palatal
(8) soft palate or
velum8/velar
(9) uvula/uvular (nyelvcsap)

1
"C" may stand for Consonants while "V" may stand for Vowels
2
The vocal tract is the cavity in human beings and in animals where sound that is produced at the sound
source (larynx in mammals; syrinx in birds) is filtered. (from wiki)
3
movements of muscles of the tongue (?), sszehzds, sszehzs
4
something which starts the sound, here, the airstram (levegramls) from the lungs
5
organs, which constribute to the process of producing sounds (wiki); 2nd terms stand for the adjective,
derivated from the noun
6
alveolar ridge is is one of the two jaw ridges (llkapocs v) either on the roof of the mouth between the upper
teeth and the hard palate or on the bottom of the mouth behind the lower teeth.
7
kemnyszjpadls, szjpadls ells rsze
8
lgyszjpadls, szjpadls hts rsze

2
(10) pharynx/pharyngeal (garat)

active:
(11) glottis/glottal (hangrs)
(12) epiglottis/epiglottal (ggefed)
(13) tongue root/radical (nyelvgyk)
(14) back of tongue body/
postero-dorsal
(15) front of tongue
body/antero-dorsal
(16) tongue blade/laminal
(17) tongue tip/apical
(18) /sub-apical (nyelvcscs alatti)
b) places of articulation9:
1. 22 bilabial (p)
2. 23 labiodental (f)
3. 162 linguo-labial
4. 163 interdental (AmE "th" )
5. 173 apical dental (BrE "th" )
6. 163/4 (laminal) denti-alveolar
7. 174 apical alveolar (t)
8. 164 laminal alveolar
9. 175 apical retroflex
10. 165 (laminal) palato-alveolar
11. 186/7 sub-apical (retroflex)
12. 156/7 palatal (Hung. "ty" c)
13. 148 velar (k)
14. 149 uvular (q)
15. 1310 pharyngeal
16. 1210 epiglottal
17. 11 glottal ()
c) manner of articulation: primarily denotes the degree of constriction
complete: stop, in which the vocal tract is blocked so that all airflow ceases.
(p, t, c, k, m, n, n, , )
almost complete: fricative, the gap is so narrow that turbulent airflow
isproduced, it has 2 types:
sibilant: (sziszeghang) (s, z, , )
nonsibilant: (f, v, , , , , x, )
almost none: approximant, the gap is so wide that there is no turbulence (w,
, , j)
9
In the list below, numbers mean: "what is placed" "where it is placed", then the name of a given sound of
that type w an example

3
none: vowel (see next time)
lateral: more constriction in the central area, more air escapes at the sides
(fricative: , ; approximant: l)
rhotic (r-sounds): trill 10(r), tap/flap 11(), approximant rhotic (, ),fricative
rhotic ()
nasal: velum is lowered, allowing air to flow through the nasal tract (as well)
(stop: m, n, , ; vowel: )
two major manner categories:
vowel vs. consonant: actual syllable peak 12vs. actual syllable edge13
(uuu=wuw, iii=jij), but the terminology is inconsistent, since syllabic
consonants are not called vowels (nnn)
sonorant (zenghang) vs. obstruent (zrelyhangok, akadlyhangok):
potential syllable peak vs. impossible syllable peak
(i) sonorants: vowels, approximants, rhotics, nasal stops
(ii) obstruents: nonnasal stops, fricatives
n.b. in many languages (e.g., English and Hungarian) obstruents come in
voiceless/voiced pairs, while sonorants are usually all voiced

d) phonation type: status of the glottis

A. glottal stop: vocal folds held together ()


B. creaky (csikorg) voice/laryngealization14:vocal folds vibrating, arytenoids
(kannaporc) pressed together, low rate of air ow ( , m, )
C. modal voice: vocal folds vibrate (b, m, )
D. breathy voice/murmur: arytenoids apart, vocal folds vibrate without contact, high
rate of air ow( / )
E. voiceless: vocal folds do not vibrate, arytenoids apart (p, m, ) (the supposed to
be under the letter)
F. aspirated15: arytenoids further apart than for voiceless (p, h)
10
In phonetics, a trill is a consonantal sound produced by vibrations between the articulator and the place of
articulation (wiki)
11
a type of consonantal sound, which is produced with a single contraction of the muscles so that one
articulator (such as the tongue) is thrown against another, so, when in e.g. in AmE, speakers say /l/ instead
of /lt/
12
a.k.a. a nucleus, a vowel, in the form of a monophthong (magnhagz), diphthong (kettshangz), or
triphthong
13
a.k.a. coda, which, comprises the consonant sounds of a syllable that follow the nucleus
14
It is a special kind of phonation, in which the arytenoid cartilages in the larynx are drawn together; as a
result, the vocal folds are compressed rather tightly, becoming relatively slack and compact.

4
e) airstream mechanism:
pulmonic 16airstream: (ex)plosive only egressive (lung air pushed out
under the control of the respiratory muscles (lgzizmok)) (p, t, k)
glottalic airstream:
egressive: (pharynx air compressed by the upward movement of the
closed glottis) ejective(mssalhangz): (p t k)
ingressive17: (downward movement of the vibrating glottis): implosive
(zrhang) (,, )

velaric airstream: only ingressive (mouth air rarefied (ritkul) by the backward
and downward movement of the tongue pressed to the velum): click (, !)
(think of Xixo (N!xau) from The Gods Must Be Crazy)
3 Multiple articulatory gestures:
a) double closures (two POAs): labio-velar stop: (kp, gb, m) or approximant: w
b) secondary articulations: labialization18: (k, g); palatalization19: (p, b);
velarization20: (m,) Englsh "dark l"; pharyngealization21: (s Arabic
languages)
c) affricates: stop+fricative: (, , , )
they are commonly sibilants (like the above)
they are in fact not doubly articulated, but slowly released stops

15
It is the strong burst of air that accompanies either the release or, in the case of preaspiration, the closure of
some obstruents.
16
related to the lungs
17
In human speech, ingressive sounds are sounds by which the airstream flows inward through the mouth or
nose (so, like taking a breath)
18
Labialized sounds involve the lips while the remainder of the oral cavity produces another sound.
19
When a consonant is produced with the tongue in a position in the mouth near the palate.
20
The back of the tongue is raised toward the velum during the articulation of the consonant.
21
The pharynx or epiglottis is constricted during the articulation of the sound.

5
6
Vowels
I Articulatory properties of vowels:
1 Properties of vowels:
a) are more difficult to characterize in articulatory terms
b) form less discrete categories than consonants
c) can partly be described as points on a continuum22
d) can partly be described in binary23 terms
2 Two kinds of properties:
a) scalar: one of a potentially infinite number of points on a continuous scale: [0,
0.1, 0.2, . . . 1]; [0, 0.01, 0.02, . . . 0.1, . . . 1]; etc.
b) binary: one of the two extreme poles (yes or no, black or white, no shades
between them): [0, 1]
3 The parameters which specify the vowel continuum:
a) height: highest point of tongue
b) frontness/backness of the highest point of the tongue
4 Binary vowel properties:
a) lip position
b) tenseness (feszessg)/laxness(lazasg)
c) nasalization see further
d) r-colouring see further
e) length
f) quality change
5 Daniel Jones (18811967): His
experiment: "A chain of small lead plates strung
together was placed on the tongue to show its outline.
The large dot added on each photograph marks the
highest point of the tongue. The cross is a point of
reference (near the end of the hard palate."
6 Tongue positions:

in the back part of


the mouth

X-ray pictures of Jones mouth in the front part of the mouth

22
Points on a continuum, i.e. different "stages" or "stops" in a continous process
23
twofold, being composed of 2 things, having 2 aspects etc.

7
diagrammes of highest tongue positions:

quite accurate representation more schematic representation

7 Lip position:
a) dichotomy24: unrounded25/spread vs. rounded
b) correlations (sszefggsek):
front vowels are unrounded, back vowels are rounded
rounded front and unrounded back vowels are rarer, and only occur if the
language has unrounded front and/or rounded back vowels, too:
front rounded V front unrounded V

back unrounded V back rounded V

the shape is read as "implies" (and pls imagine that it's like a letter
U, that was the best I could come up w....)

8 The Jones vowel chart:


notes:
close = high ( Hung. magas
magnhagz)
open = low
front = palatal (= Hung. magas)
ack = velar (= Hung. mly)
the vowel is called schwa
the vowel is called ash

24
A dichotomy is any splitting of a whole into exactly two non-overlapping parts, meaning it is a procedure in
which a whole is divided into two parts.
25
In phonetics, vowel roundedness refers to the amount of rounding in the lips during the articulation of a
vowel. That is, it is vocalic labialization. When pronouncing a rounded vowel, the lips form a circular opening,
while unrounded vowels (also called spread vowels) are pronounced with the lips relaxed.

8
9 Cardinal vowels: They are arbitrarily chosen points in the vowel chart, to which
vowels of specific languages can be compared.

primary cardinal Vs secondary cardinal Vs

NB: e.g., i is primary cardinal


vowel number 1; o is primary
cardinal vowel number 7; is
secondary cardinal vowel
number 2; etc.

cardinal V positions

an alternative analysis for vowel height:

tense vowels: close and close mid lax vowels: near close and open mid
vowels vowels

(a) distinguishing tense and lax vowels: lets us reduce the five vowel heights (close,
near close, close mid, open mid, open) to three (close, mid, open): close is now
close tense; near close is close lax; close mid is mid tense; open mid is mid lax

9
10 Further vowel contrasts:
a) nasalization: the velum is lowered, air flows out both through the oral and the
nasal tract (mouth and nose)
b) nasalized vowels tend to lower (become more open), cf. French:
finir [fini] finish vs. fin [f26] end
c) r-colouring:
the tongue tip is curled back behind the upper alveolar ridge
r-coloured (retroflex or rhotic) vowels are quite rare, but some varieties of
English have them, e.g., in some AmE:
merger [m]
d) quantity:
vowel length:
It is also a binary property, i.e., a vowel is short or long
is marked by two triangles facing each other: (can be replaced by a
colon)
the quality of a long vowel:
may be stable long vowel
may change diphthong
e) diphthongs:
types of diphthong:
height of the two parts:
closing: ai, au (second half closer than first)
centring:a, i, u (second half central)
opening: ua, ia (second half more open than first)
syllabicity:
falling27: ui (syllabic first half)
rising28: ui (syllabic second half)
the analysis: of diphthongs is language specific: they may be analysed as
consonant+vowel or vowel+consonant sequences as well

26
please imagine that it has a tilde (~) on the top, not '-smth mark
27
Falling (or descending) diphthongs start with a vowel quality of higher prominence (higher pitch or volume)
and end in a semivowel with less prominence
28
Rising (or ascending) diphthongs begin with a less prominent semivowel and end with a more prominent full
vowel

10
Acoustic phonetics
I Phoneticians: Experts who study sounds in the respect of human beings. It has 2
branches:
1 Articulatory phoneticians: Study the organs, which produce human sounds (e.g.
glottis etc.) and exactly how humans produce sounds. Therefore, articulation is
based on articulators (organs to produce sounds) and mostly on human anatomy.
2 Acoustic phoneticians: Study the sound waves, through which sounds are
transmitted. Therefore, it deals with the perception of sounds, audition (halls mint
kpessg) and properties of speech sound waves.
II Sounds:
1 Sound transmission: If there's a sound, there's a sound transmission (hangtvitel?) It
has 3 components:
a) sound source: producing sounds
b) transmission through a medium (e.g., air, water): how they spread
c) potential29 receiver/interpreter: One who perceives the sounds and interprets
(rtelmez) it
2 Definition of sounds: Sound is an audible disturbance (hallhat zavar?) of a medium
produced by a vibrating30 source (e.g. vibration of the air particles).
3 Measuing sounds: There are 2 problems with it: sounds are invisible and most of them are
fairly complex; therefore, sounds have to be made visible for an analysis and first, simpler
sounds have to be dealt with.
4 The simplest sounds: They are called pure tones (e.g. the tuning fork
(hangvilla) emits pure tone31)

29
lehetsges, as opposed to actual, who is doing it just now, not in the future, not in the past
30 the sound of the tuning fork
rezg, if something vibrates, it produces sounds
31
the sound depends on the length of the 2 prongs (g, villa)

11
a) basic terms:
frequency: movement of the
tuning fork
(a) a complete movement:
starting/rest position
maximum displacement
(elmozduls) back over starting position
maximum displacement back to
starting position a cycle (it's
sign: c)

(b) frequency (F): the number


of completed cycles per
second (s) (Hertz (Hz) or cps), the
tines (villk) complete 440
cycles per second, so the
frequency of the tuning fork is 440 Hz

simple periodic motion (SPM):
(a) the SPM of the tuning fork:
the tines of the tuning fork vibrate in simple
periodic motion; simple, because the tines move back/forth one fixed
number of times per second (no matter how hard the fork is struck); and it has
periodic motion, because the pattern repeats itself until it damps out (so, it's like
a swing (hinta), only a bit more complex)

b) graphing SPM: with the help of a waveform/oscillogram, which is a graphical


display of how amplitude varies over time.

with orange circles, we can see the amplitude peaks, while the line forming pits and hills is called a sinus waveform

waveform: It consists of two independent properties.

12
(a) amplitude: distance from zero crossing, amplitude peak is the highest amplitude
value
(b) time: there are 2 ways of talking about time in the
waveform
1. expressed as period (T) = time for a cycle
to complete, unit: seconds T = 0.01 s
2. expressed as frequency (F) = number of cycles in a second, unit: Hz
F = 100 Hz

same amplitude, different frequency same frequency, different amplitude

c) pressure-based graph: it shows the variations in air pressure with respect to an


equilibrium (egyensly?), the continuous line here is called the sinusoid
waveform32
d) change in frequency: su jective sensation (benyoms) of pitch33

32
szinuszhullm
33
hangmagassg

13
e) change in amplitude: su jective sensation of loudness/intensity

f) basic properties of simple sound waves:

Acoustic property Unit Perception


frequency Hz pitch/intonation34
amplitude decibel (dB) loudness/intensity
the threshold of hearing (0 dB = 20 Pa): the buzz of a mosquito around 3 meters
away
average street traffic: 80 dB (t 100000 Pa)
threshold of pain 140 dB (100,000,000 Pa): jet engine at 25m distance

2 Complex sounds: They are nothing like pure sounds


a) characterising complex waves: Based on the theory of Jean Baptiste Joseph Fourier
(17681830). Its key idea is the theory that if we can reduce a complex periodic
waveform into a combination of sine (szinusz) waves, then we can describe it using
information about the frequency and amplitude of each component sine wave.
(e.g. 1 100 Hz/30 dB sinus wave + 1 200 Hz/10 dB sinus wave + 1 300 Hz/20 dB sinus wave)
this is called a synthesis: the addition of two or more different sine waves to create a
complex periodic wave

34
hanglejts

14
b) properties of complex periodic waves:
the amplitudes of the complex periodic wave depends on
the addition of the amplitudes of the component waves
the frequency of the complex periodic wave depends on
the frequency of the sine wave with the smallest
frequency
we call this the fundamental frequency (f0) of the
complex periodic wave
the other sine wave frequencies within the complex
periodic wave are called harmonics (H) (or: overtones
felhangok)
c) harmonic series35: Our example periodic complex wave has
this harmonic series (also called Fourier series):

Harmonic Frequency Amplitude fundamental frequency of a complex


periodic wave
H1 (= f0) 100 (100 1) 30 dB

H2 200 (100 2) 10 dB

H3 300 (100 3) 20 dB
harmonic analysis

d) harmonic analysis: the reverse of


synthesis, finding (characterizing) the
component sine wave, harmonics of the
complex wave
e) Fouriers theorem (1822):
All complex periodic waveforms
can be analysed into a sum of
sinusoidal component
waveforms (harmonics)36
The mathematical algorithm of this process of harmonic analysis is called
Fourier analysis or Fourier transformation.

35
A Fourier (or harmonic) series decomposes periodic functions or periodic signals into the sum of a (possibly
infinite) set of simple oscillating (rezg) functions, namely sines (sin) and cosines (cos), it is originated from and
mainly used in mathematics (and this is the point, when we DON'T want to know more abt it...:D)
36
A sum of sinusoidal component waveforms (szinuszalak komponenshullmok sszege (slogan of the
lecture...)), so every complex wave can be decomposed into its sinus-shaped components

15
f) graphical representation of harmonic analysis: spectrum graphs, it is like the
spectrum of light (e.g. a prism)
the sound spectrum: (plural: spectra) It is a plot of the results of harmonic
analysis, it is the basis of sound quality/timbre (hangszn)
frequency of harmonic: horizontal axis
amplitude of harmonic: vertical axis
time: not shown (Fourier analysis is taken at a particular instant of time
g) summary: the basic acoustic
aspects of sounds
all these components are independent
of each other: sound loudness
depends on amplitude
sound pitch depends on f0
sound quality/timbre depends on
the spectrum (harmonic series)
a contrast: depends on the
spectrum
e.g. ead / id/ arred /b d/

3 Sound types: sound types based on


acoustics
a) the prototypical aperiodic sound:
white noise37

37
The sound produced by e.g. the radio, when it is tuned between the channels, quasi a crepitation (sercegs)

16
II Producing speech sounds: (thinking of a simple experiment38)
1 Resonators39 and filters40: the sound of the resonating object largely depends on its
shape and size:
a) air vibration and size of resonator:
small object: quick vi ration higher frequency (so higher pitch)
large object: slow vi ration lower frequency (so lower pitch)
b) sourcefilter model: what do
resonating filters do?

c) the basic acoustics of periodic sounds: the sourcefilter


model of periodic speech sounds (e.g., vowels)
production of periodic speech sounds: it can be
modelled with the input source + resonator filter
model
source: the periodic vibrations generated by the
vocal folds in the larynx
filter: the vocal tract cavities (pharynx, oral cavity,
and nasal cavity), which modify the shape of the
input source
output: the actual speech sound uttered

the sound [3:] in the word bird

38
Put an empty bottle near your mouth with its hole towards your lips holding it vertically and try to blow the
air over it. If you succeed, you will hear a whistle-like sound, produced by the process of filling the empty bottle
with air.
39
A resonator is a device or system that exhibits resonance or resonant behaviour, that is, it naturally oscillates
at some frequencies (http://en.wikipedia.org/wiki/Resonator)
40
A device for isolating a certain frequency band from a complex sound.
(http://encyclopedia2.thefreedictionary.com/Filter,+Acoustic)

17
2 Formants (formns): formants are frequency peaks in the output signal; they arise
because the vocal tract modifies the vibrations of the vocal folds (hangredk?)
a) some basic information on formants:
formant frequencies depend on the shape & size of the speech tract
size/shape can depend on the position of articulators, like the tongue
frequency of the first three formants (F1, F2, F3) are the most important for
setting phonetic quality and phonological contrast
F1 pharynx/throat; F2 oral cavity
e.g., the vowel in bird [3:] can be characterized by having formants at: F1 =
500 Hz, F2 = 1500 Hz, and F3 = 2500 Hz
b) correspondence between tongue position and formants:

front vowels have


high F2
low vowels have
high F1
rounding generally
lowers
formants

3 Representing spectra in "real time": by the use of spectrograms they are series of
spectra in quasi-3D
a) they encode information on:
time: (in seconds) axis x
frequency: (in Hz) axis y
amplitude: (in dB)
greyscale

18
spectrograms
of bead & barred

19
The consonant phonemes of English
I Basics of the phonemic analysis: (as to make further examples)
1 The sound [] and [n] in British Accents: Both Southern and Northern English has this
sound; however, there's a difference.
a) Northern English: e.g. in the word thing [g]
b) Southern English: e.g. in the word thing []
c) difference: the 2 accents deal with the sound [n] and [] differently
NBE: thin [n] vs. thing [g]
SBE: thin [n] (same) vs. thing []
so: the [n] vs. [] contrast distinguishes words in SBE, but does not
distinguish anything in NBE: pronouncing [n] instead of [] would cause no
confusion, since the sound [g] is pronunced
2 Terminology:
a) minimal pair: two words (i.e., with different meaning) of equal length that differ
in one sound only, e.g., thing and thin in SBE ([n] and [g])
b) free variation: a word with several possible pronunciations (used by natives
sometimes in the same accent based on an arbitrary choice)
e.g., unkind [nkand] and [nkad], either [a] and [i]
c) environment: the context in which a (linguistic) element occurs,
e.g., [_] is an environment of [] and [n] in SBE
[_kand] is an environment of [] and [n]
[_] is an environment of [a] and [i:]
d) distribution (elolszls): the set of all the evironments of an element, a long list,
but it can be generalized: instead of [_], [s] etc., we can say V_#
(after a vowel, at the end of the word)
complementary distribution: two distributions are complementary if none of
their elements are common i.e., they do not overlap
e.g., A = {a, b, c},B = {d, e, f } (or like the distribution of the indefinite article a
and an)
overlapping distribution: two distributions are overlapping if some (or all) of
their elements are common, i.e., they are not complementary
e.g.: A = {a, b, c}, B1 = {c, d, e}, B2 = {b, c}, B3 = {a, b, c}
the distri ution of [] in NBE:

that is, [] occurs before [k] or [g] only


e.g., ink [k], king [k]

20
the distribution of [n] in NBE:

that is, [n] occurs before [t] or [d] or # (word boundary, i.e., at
the end of the word) or in several other environments (e.g.,
before vowels, [s], [r], [l], etc.), but not before [k] or [g], e.g.,
tent, bend, ten, name, tense, Henry, only

the distribution of [n] and [] in NBE: is complementary, the two distributions


do not overlap: [] occurs only where [n] does not occur (or if they do cooccur
(e.g., in unkind) theres free variation)
the distribution of [] in SBE:

that is, [] occurs before [k] or [g] or #


e.g., ink [k], younger [jg], young [j]

the distribution of [n] in SBE:

that is, [n] occurs before [t] or [d] or # or in several other


environments, just like in NBE

the distri ution of [n] and [] in SBE: is not complementary, the two
distri utions overlap, oth [n] and [] occur word finally ( _#)
3 Allophones (alakvltozat) of a phoneme:
a) if two elements are in complementary distribution:
they cannot contrast41
there exist no minimal pairs (e.g., with [n] vs. [] in NBE)42
their difference has no function
their difference goes unrecognized by native speakers
they are considered the same sound y native speakers

41
Two language-elements are in contrast in a particular unit if replacing one
by the other changes the meaning of the unit, so they do not distinguish 2 separate sounds, so all the variants
of the phoneme is thought to be the same, differing only in pronunciation ([]=[n])
42
so, in NBE, it doesn't matter whether we have [n] or [], they are considered as just the same sounds

21
it can usually be predicted which of the two occurs in a given environment
(e.g., efore [k] or [g], it will e [], elsewhere [n])
two such sounds: are allophones of the same phoneme; [n] and [] are
allophones of /n/43 (or / / or //, the symbol is arbitrary); this is true for
NBE, not necessarily in any other language (incidentally it is also true for
Hungarian)
4 Jan Ignacy Niecis law Baudouin de Courtenay44 (18451929): (a.k.a. Ivan
Aleksandrovich Boduen de Kurtene) He is the father of the modern theory of
phonemes.
5 Separate phonemes:
a) if two elements are in overlapping distribution:
there exist environments in which both occur (both [n] and [] occur in _#)
they may contrast (if we dont have free variation), if they do:
(a) there exist minimal pairs (e.g., thin /n/ and thing //)
(b) the difference has a function (e.g., distinguishing thin and thing)
which of the two occurs in a given environment is often unpredictable: one
cannot tell whether /n/ or // occurs in the environment /_ /; this is a
lexical property of the word (it is an irregularity)
two such sounds: are two separate phonemes; /n/ and // are two separate
phonemes in SBE
6 Weaknesses of the classical phonemic analysis:
a) in SBE /n/ and //:
do not contrast in most environments: [_k], [te_t], [_et]; because the
distri ution of // is defective45
in these environments the occurrence of /n/ vs. // can e predicted
b) are they really phonemes then?: this is a necessity in taxonomic (rendszertani)
analysis, summarized as once a phoneme always a phoneme46
c) // and /h/ are in complementary distri ution, too: so why arent they
allophones of a phoneme? esides complementary distri ution, phonetic
similarity (e.g. /n/ and // just sounds to e similar for some cause, while // and
/h/ just do not) is also necessary for establishing a phoneme.

43
when we talk about phonemes, we use //, when allophones of phonemes, we use []
44
a Polish person of French origin, who migrated to Russia to study English...ok...
45
limited, restricted; they are absent from some positions (initial, medial kzpen van and word-final) in
words
46
If they have a distribution in any kind of accent in which they are different phonemes (SBE), they are also
considered to be phonemes in any distribution; alothough, they may behave like allophones (NBE)

22
7 Transcriptions: two kinds of transcription
a) narrow/phonetic transcription47:
includes subphonemic detail, predictable information48, allophones
usually concentrates on a particular aspect of the pronunciation
totally narrow transcription is impossible (since every person speaks in a
slightly different way)
is enclosed in brackets (square brackets)
e.g. peeled [pid]49
b) broad/phonemic/phonological transcription50:
includes only phonemes, unpredictable information
is enclosed by slashes (a.k.a. virgules)
e.g. peeled /pi:ld/, thing SBE //, NBE /g/ but think is /k/ (not /nk/)
in SBE, although it predictably has //:

II Consonant phonemes of Standard Southern British English:


1 Obstruents:
a) stops (explosives)51:
labial (bilabial)52: p - b (pan - ban)
coronal53 (alveolar54): t - d (tan - Dan)
velar55: k - g (con - gone)
voiceless (aspirated): p - t - k (pool - tool - cool)
voiced (unaspirated): (b - d - g) (bay - day - gay)
affricates56: - (choke - joke)
b) fricatives:
labial (labiodental57):f - v: (fan - van)
dental (foghang): (thigh - thy)
alveolar58:s - z (seal - zeal)

47
BEWARE!!! This is not the one that we call in Hungarian fonetikus trs and not the one, which is commonly
refered to as phonetic transcription and used in primary and grammar schools!!!
48
use of different things e.g. aspiration, flapping, nasalisation or palatalisation, which are predictable being
specific to a particular geographic location, social class or usual in a particular environment in words
49
aspiration, it is a strong burst of air that accompanies either the release or, in the case of preaspiration,
the closure of some obstruents; "dark l", the sound [] is used in the word Hell, and in most American
accents
50
this is the one called "phonetic" in non-academic level
51
the vocal tract is blocked so that the airflow ceases then it is opened as a thin gap when the air escapes like
an explosion
52
articulated with both lips
53
those produced using the tongue tip or blade, and typically alveolars; used as a collective adjective for
dentals, alveolars and palat-alveolars
54
articulated with the tongue against or close to the superior alveolar ridge
55
Articulated with the back part of the tongue against the soft palate
56
Affricates are consonants that begin as stops but release as a fricative (rshang).
57
Labiodentals are consonants articulated with the lower lip and the upper teeth.
58
Articulated with the tongue against or close to the superior alveolar ridge.

23
postalveolar59: - (Aleutian - allusion)
voiceless: (fie - thigh - sigh - shy)
voiced: (breve - breathe - breeze)
2 Sonorants (zenghang):
a) nasals: m - n - (Pam - pan - pang), p - b - m (pie - bye - my), t - d - n (ton - done -
nun), k - g - (rack - rag - rang)
b) nonnasals: l - r - w - j - h (lot - rot - what - yacht - hot)
III non-SSBE:
1 [] (= SSBE [w] so the sound [m]!!!): w - (witch - which, wear - where, Y -
why)
2 [x] (= SSBE [k] so not an x!!!): k - g - x (lock - log - loch)

59
Articulated with the tongue near or touching the back of the alveolar ridge.

24
The vowel phonemes of English
I Some clarification...:
1 "Englishes": what does English mean?
English, like any other language name, refers to an a straction (elvont dolog); a
question like does English have the phoneme //? is not meaningful (since the
language is a whole bulk of accents (akcentusok sszessge) and varieties); unless,
Eng is considered as a shorthand ("rvidts") for a given accent of English; in this
course, it is a shorthand for Standard Southern British English (SSBE), a.k.a. Received
Pronunciation (RP) (used less and less frequently), BBC English, Queens/Kings
English.
Accents of English are much more diverse in their vowel than in their consonant
inventories (mssalhangz kszlet)
II Selecting symbols:
1 Analysis in transcription: Vs can be analysed in 2 ways concerning transcription
(difference between bead and bid will be used for demonstration)
a) the difference can be seen...:
as a quantity60 difference: long vs. short
as a quality61 difference: tense (feszes) vs. lax (laza)
b) accordingly, the two vowels may be transcribed as...:
long-short: /i:/ /i/ (e.g., Jones English Pronouncing Dictionary, old Orszgh
dictionaries)
tense-lax: /i/// (e.g., Kenyon & Knotts A Pronouncing Dictionary of
American English, Giegerichs English Phonology, etc.) (NOTE!: ordinary "i" vs.
"i" without a dot!!!)
combination of the two: /i:/// (e.g., Gimsons, Wells pronouncing
dictionaries, new Orszgh dictionaries, most current British dictionaries,
Professor Ndasdy (), and also this course
2 Phonetically:
a) /i:/ is always tense, // is always lax
b) /i:/ is sometimes long (bead /bi:d/), sometimes short (beat /bi:t/)62
c) /I/ is shorter in pick than in pig
(note that the broad transcription does not indicate b and c)
d) conclusion: the symbol pair /i/ // is the most appropriate for this contrast,
since tense/lax is a constant phonetic property of this opposition, long/short is

60
hangmennyisg, how much "time" does it take for the speaker to produce the sound
61
hang jellege, how much does the speaker have to strain his/her articulatory muscles (muscless of the mouth,
jaws etc.)
62
I assume that bead and beat are both longish, but bead is even longer in pronunciation (comp bead, beat vs.
bit), but I'm unsure about this

25
variable; nevertheless, the symbol pair /i:/ // is more wide-spread, since it is
more contrastive visually.

III SSBE vowel phonemes:


1 Monophthongs63:
accidental gap: it is accidental that
there is no word /bt/ (and /bt/) in
SSBE, but other pairs prove that the
difference in the vowel in but and put is
not due to the voicing of the first
consonant: bushpush, bunkpunk
alternative symbols: /e/ is sometimes
transcribed as //, // as /a/, not in this
course!

the high long monophthongs: may diphthongize64, especially word finally65 and
prevocalically66: who, me? [hu mi] (=/hu: mi:/)
2 Diphthongs:
closing diphthongs67
centring diphthongs 68
sound //69: contrasts for //
quoit // - quote /o/ - quite /a/

sound //: it is obsolete, it is


replaced by /:/

alternative symbols:
/e/ is transcribed // by Jones and McMahon

63
sometimes, they are translated as magnhagzk, but it is by no means true in Eng, since not every vowel is a
monophthong, we hear only 1 "kind of" sound when it's produced (unlike in standard HUNG, in which we have
monophthongs only)
64
pronunced as a diphthong
65
adj, at the end of the word
66
concerning the initial sound segment of vowels
67
closing dipthtongs: the second element is more close than the first
68
centring vowels: A centering diphthong is one that begins with a more peripheral vowel and ends with a
more central one
69
it souns like the German word neun (nine)

26
/o/ is transcribed // by Gimson
the true diphthongs: (a.k.a. wide or low diphtongs)
/a/ // /a/

3 Other vowels of SSBE:


a) triphthongs (hrmashangzk): the true diphthongs (/a a/) may occur
followed by a schwa (//); some analyse these as triphthongs: fire /fa/, power
/pa/, Moir /m/; others analyse them as heterosyllabic (belonging to two
syllables): /fa./

b) monophthongization70/smoothing: centring diphthongs and triphthongs tend


to monophthongize:

this process causes merger71 in cases i, ii, and iv, resulting in homonyms, but
not in case iii
c) sound /ju:/:
some analysts consider /ju:/ and /j/ unitary (egysges) phonemes, a rising
diphthong and a triphthong, respectively
others treat these as consonant + vowel sequences
d) unstressed syllables:
typically contain // (schwa)
may also contain /I i ju: ju j j/ or a sonorant consonant
in SSBE schwa occurs exclusively in unstressed syllables

70
when a sound is written as a diphtong but pronounced as a monophtong
71
Phonemic merger is a loss of distinction between phonemes.

27
4 Standard lexical sets: A Standard Lexical set is a group of words represented by a key
word (set in SMALL CAPITALS) containing the same vowel phoneme. In a given
accent, several lexical sets may contain the same vowel; in other accents, they may
contain different vowels. The system was devised by John Wells.

a) realizational difference72: It is linguistically irrelevant because it does not affect


the linguistic system. If the knight 73is replaced by another object in a game of
chess, that object can also function as a knight, sais Ferdinand de Saussure (NB!: a
thing must stand for only one thing74)

b) SSBE vs. New Zealand English:


SSBE has the same vowel in dress as NZE
in trap, but the contrast between dress
and trap remains in both languages

c) systemic difference: it is linguistically relevant, it affects the linguistic system. It is


a contrast that exists in one accent but absent from the other.
SSBE vs. Singapore English:

SgE has the same vowel in dress and


trap bat and bet are homonyms in
SgE, but a minimal pair in SSBE

72
difference in pronunciation, simply put, the statement means that if we take the word but, it doesn't matter
whether we pronounce it as /bt/ or/bt/ because we mean the same word, it doesn't have a different
meaning
73
famous theoretical experiment of Saussure
74
so natives are unlikely to say /bt/ on one occassion and /bt/ on another

28
SSBE vs. General American:

in SSBE thought, north, and force have


the same vowel (force had // before it
monophtongized) GA has three
different vowels: morning and mourning
are homonyms in SSBE, but a minimal
pair in GA

5 Charting vowel phonemes: SSBE vowel phonemes in the Jones vowel chart

29
The R rules
I Grouping vowels: different groupings
1 First kind of grouping:
a) occur in a stressed syllable: (call it "group (A) now) /i: e e e a a a: : o
: u: /
b) do not occur word finally: (B) / e /
c) occur before another vowel: (C) /i: e a a o u: /
d) do not occur before /r/ within a morpheme: (D) /i: e o u: /
e) do not occur before a noncoronal75 consonant cluster (e.g., /sk mp lp kt p/): (E)
/i: e a o a o u: e : /
2 Second kind of grouping:
a) set 1: (C) /i: e a a o u:/ i.e. the noncentring76 diphthongs and the
potentially diphthongal long monophthongs77 [i u]
b) set 2: E\C / e :/ i.e., centring diphthongs and /:/ the FORCE vowel,
which monophthongized from []78
c) set 3: (B) / e /, i.e., short vowels
d) set 4: A \ (B E) / : : :/ i.e., some long monophthongs /:/ is the vowel
used in morphemes THOUGHT and NORTH
3 Vowel categories: different labels are given to the groups established before
a) tense vowels:
plain tense vowels: set 1 /i: e a a o u:/
broken tense vowels: set 2 / e :/
tense vowels are marked with a macron (the line on the top of the V):
tense vowels are also referred to as long, or free
b) lax vowels:
plain lax vowels: set 3 / e /, marked with a breve:
broad (lax) vowels: set 4 / : : :/, marked with a circumflex:
lax vowels are also referred to as short, or checked

75
The word denotes dental, alveolar and palato-alveolar sounds. Coronal consonants are consonants
articulated with the flexible front part of the tongue. Only the coronal consonants can be divided into apical
(using the tip of the tongue), laminal (using the blade of the tongue), domed (with the tongue bunched up), or
subapical (using the underside of the tongue), as well as a few rarer orientations, because only the front of the
tongue has such dexterity. see more at: http://en.wikipedia.org/wiki/Coronal_consonant
76
A centring diphthong is one that begins with a more peripheral vowel and ends with a more central one,
think of the cardinal V chart
77
see Vowel phonemes of English "who, me?" example
78
it used to be /fs/ but today it's pronounced as /fs/

30
II Rhotic and nonrhotic accents: Generally speaking, British accents are rather nonrhotic
than rhotic while American ones are vica versa; however, there are exceptions in both
"groups" of accents.
1 SSBE79 is a nonrhotic accent: that is, not all orthographic/historical80 Rs are
pronounced: no /r/ is pronounced in part /p :t/, beard / d/, martyr /m :t/,
merger /m:/, iron /an/, etc.
2 In a rhotic accent (e.g. GA, General American): the /r/ (or its colouring81 on the
preceding vowel) is preserved: part /p rt/ or /pt/, etc.
3 The distribution of /r/ in SSBE: SSBE is a typical nonrhotic accent, because /r/ occurs
only before vowels (and syllabic consonants82), elsewhere it is dropped

a) accordingly:
/r/ is dropped at the end of words (e.g., car /k :/), but not if the next word
begins with a vowel (e.g., car engine /k :r enn/) such an R is called a
linking R
an /r/ is also pronounced after any word-final83 / : : / if the next word
begins with a vowel: e.g., Shah of Persia / :r v p:/, law and order (!!!)
/l:r nd :d/, vanilla ice (!!!) /vnlr as/ such an R is called an
intrusive R
b) linking and intrusive R: they are not any different linguistically: both Shah and star
end in a vowel in isolation (/ :/, /st :/)
end in a vowel when followed by a consonant (the Shahs name /- :z-/; the
stars name /-st :z-/)
end in /r/ when followed by a vowel (the Shah of Persia /- :r v -/; the star
of Persia /-st :r v-/)
but: a dilemma!!!
do both Shah and star end in /- :/ with /r/ inserted before vowels?
or
do both end in /- :r/ with /r/ dropped when nothing or a consonant follows?

79
Standard Southern British English
80
those, which are written down
81
Simply put, the vowel assimilates to the sound /r/, to hear some, go to: http://en.wikipedia.org/wiki/R-
colouring
82
A syllabic consonant is a consonant which either forms a syllable on its own, or is the nucleus of a syllable.
Many dialects of English may use syllabic consonants in words such as rhythm /rm/, but other experts claim
it's rather a schwa () and the exact consonant /rm/
83
simply, when the word ends with those vowels like these

31
c) vowels in the vicinity (krnyezet) of R: (position 1 and 2 are of course not names
:D I just use them here)
in position 1: r_ (i.e., after /r/) any vowel may occur, e.g.
read /i:/, raid /e/, ride /a/, Roy //, road /o/, root /u:/, etc.
rear //, rare /e/, rural //, roar /:/, etc.
rid //, red /e/, rat //, rum //, rot //, rook //, etc.
raft / :/, raw /:/, etc.
in position 2: _R (i.e., before an orthographic R, pronounced or not) the
distribution of vowels is more restricted, the choice depends on:
(a) the type of vowel: tense (vs. true diphthong) vs. lax
(b) the position of the R: tautosyllabic84 or heterosyllabic85, tautomorphemic
or heteromorphemic
d) tense vowels before R:
plain tense vowels: generally do not occur before tautomorphemic R each
plain tense vowel has a broken tense counterpart, which replaces it in the
pre-R environment. True diphthongs retain their plain version before a
heterosyllabic R in present-day SSBE

e) lax vowels before R:

plain lax
vowels: do not occur
before tautosyllabic R
# (word boundary) is not
always indicated in
spelling: (* means an ill-
formed smth)
pay#roll (cf. /perol/,

84
the /r/ is in the same syllable, same for morphemes
85
the /r/ is in another syllable, same for morphemes

32
*/perol/)
furr#y /f:ri/, */fri/ vs. hurry /hri/, */h:ri/ (but cf. GA /hri/)

f) pre-R vowels not followed y R:


roken vowels86 not followed by R:
museum /mjzm/, fuel /fjl/, etc. seem to have a broken vowel without a
following R; in fact, we witness vowel coalescence/contraction
(egybeolvads/sszehzs?) here:
/mj zi: m/ /mj zm/
/fju: l/ /fjl/ (it is not */fj:l/, it shows that this is not the usual broken
vowel, cf. poor /p/ /p:/)
broad vowels not followed by R:
ask / :sk/, dance /d :ns/, path /p :/, tomato /tm :to/, schwa /w :/,
colonel /k:nl/, laud /l:d/, hawk /h:k/
/:/ is the most common, typically spelt au, aw, or ough
/ :/ in many of the examples is a SBE phenomenon, here other accents
have //
/:/ is exceptional
III L breaking , broadening88, dropping89: with less consistency, L exhibits properties similar
87

to those of R

86
In historical linguistics, vowel breaking (sometimes called vowel fracture) is the change of a monophthong
into a diphthong or triphthong. The change into a diphthong is also known as diphthongization. Vowel breaking
is often distinguished from diphthongization and defined more narrowly as a harmonic (i.e., assimilatory)
process involving diphthongization triggered by a following vowel or consonant. see more at:
http://en.wikipedia.org/wiki/Broken_vowel
87
L-breaking is a process occurring in Modern English in which historical /l/ becomes syllabic /l/ (so constitutes
another syllable!!!) after certain vowels. L-breaking occurs generally after the diphthongs /a/, // and /a/. As
a result, historically monosyllabic (1 syllable) tile, boil and fowl come to rhyme with historically disyllabic (2
syllables) dial, royal and vowel. L-breaking is also common in rhotic varieties of English, after // and //, hence
pronunciations like /wld/ for world.
88
. Simply put, the vowel after which the /l/ stands broadens, and the orthographic "l" is present in
pronunciation, too
89
. Simply put, the vowel after which the /l/ stands broadens, while the orthographic "l" is ignored in
pronunciation

33
Sound values of single vowel letters and
vowel digraphs
I Spelling: the most widespread transcription convention of a language is called its
orthography or spelling
1 Conservative aspect: it is often much more conservative than the language system;
therefore, its relation to phonemic transcriptions is not one-to-one
a) homophones90:the same phoneme can be spelt in different ways, e.g., sawed,
soared and sword are all /s:d/ such words are homophones
b) homographs91: the same letter (combination) can represent different phonemes,
e.g, EA is /i:/ in lead vezet, /e/ in lead lom such words are
c) homonyms92: e.g. led and LED are of this kind
2 From a linguistic point of view, spelling is irrelevant, since:
a) it is an often inconsistent convention
b) it does not have any influence on the linguistic system (or very little)
c) it is not a part of the linguistic knowledge of many speakers, who nevertheless
speak the language perfectly well (little children, illiterate people, speakers of
languages without an orthography
3 Spelling and learning: spelling is relevant for learners of English as a second
language, since they typically meet new words in writing first

90
sounds as the same
91
written as the same
92
both written and pronounced as the same, NB! not to be confised with polisemy, the meaning of one lexeme
is metaphorically extended on the basis of some similarity, cf. leg (of a man) vs. leg (of a table)

34
II Spelling of consonants: (here, we deal w 2 rules ONLY, but there are more of them of
course)
1 The letter C has two basic sound values93:
this is a spelling rule: it does
not matter what the vowel is
pronounced, or whether it is
pronounced at all: in cater,
cure there is a front vowel/glide after the letter C, in cite there is a back vowel after
it, in nice there is no vowel pronounced after it, note Celtic /keltIk/ but /seltIk/ as a
football team
2 The letter G has two basic sound values:

but it is more complicated: gaol /eIl/, gynecology /gankli/

3 Sound values of single vowel letters94:


the sound value of I
and Y is equivalent
the plain tense
value is the name of
the letter; /a/ and
// do not have any
single vowel letter
spelling
// is missing from
the plain lax values95
E, I, and U have the
same broad lax value
(first, NURSE, merger)
O has the same broken
tense and broad lax
equivalent(smoothing
of []

93
more or less hangrtk, the actual pronunciation
94
broken tense vowel letters are those dipthtoungues and triphtoungues which are "broken" or "curved" by a
schwa //, with the exception of /:/. English has actually 3 of "this" sound (here, broaken tense /:/ refers to a
sound thad was prononunced as //, but no longer), it is in SSBE that they pronounced as the same, American
accents usually distinguishes all of them, to see more, check The background for Engish Pronunciation y dm
Ndasdy
95
This vowel has no regular spelling equivalent, it is usually spelt with u (full) or oo (book)

35
4 Vowel digraph: It is a combination of vowel letters which represent one phoneme (cf.
consonant digraphs in E: ch, sh, th; and Hu: cs, gy, ny, sz, etc or traditional vowel
digraphs in Hu names: Pal, Ver, Ers, Etvs, Thewrewktrk)
a) NB: some sequences of vowel letters are not digraphs
in quit UI represents /w/ followed y //
in opaque U obligatorily follows Q, E is mute
in buy, build, circuit, biscuit, guy, guide U is mute
in die mute E follows the single vowel letter I (cf. dying) an English rule:
smallest permitted words must be composed of at least 3 letters
in mosaic, create, reinforce, neon, museum, client, boa, heroin, cooperate,
ruin the underlined vowels represent two separate vowels, according to an
old-fashioned convention, some of these are spelt with diaeresis: coperate,
nave
note, however, zoological (gardens)96, originally with /o/, later with /u:/,
clipped97 to zoo (spelling influences the language here)
b) sound values of vowel digraphs:

96
Zoological garden was the full name of what we call today zoo.
97
Clipping means shortening a lexeme and thus producing a more informal
variant, e.g. PHOTOGRAPH PHOTO

36
c) quality deviations:

37
38
39
40
Graphic positions & overriding rules
I Predicting the sound value of single vowel letters:
1 Single vowel letters: regularly have four possible sound values in stressed syllables
a) plain tense: e.g., /e/
b) broken tense: e.g., /e/
c) plain lax: e.g., //
d) broad lax: e.g., / :/
2 The choice is often predictable from the spelling:
a) plain vs. broken tense: no R, heteromorphemic98 R vs. tautomorphemic99 R (e.g.,
cane, haze, payroll with /e/ vs. care, hare, airy with /e/); hetero- vs.
tautosyllabic R for true diphthongs100 (virus vs. wire)
b) plain vs. broad lax: no R, heterosyllabic R vs. tautosyllabic R (e.g., can, has, barrel
with // vs. car, hard, barred with / :/)
c) tense vs. lax: graphic position, overriding laxing rules
3 Graphic positions:
a) free position:
V spa, she, ski, sky, pro, flu
VV chaos, neon, trial, dryad, stoic, fluid (not digraphs!)
VCV stake, serum, writer, Skype, glory, flute
VCOCL acre, secret, trifle, Cyprus, ogre, scruple
CO = obstruent (P, T, C, B, D, G, F), CL = liquid (L, R)
b) covered position:
VC map, pet, big, gym, for, pub
VCC hand, pester, miss, myrtle, worse, hurry
c) discrepancies (ellentmondsok):
consonant digraphs count as one C: bathe, hyphen, ache free position
the letter X counts as two Cs: taxi, Texas, vixen, oxen, buxom covered
position
4 The basic rules:
a) in free position single vowel letters are pronounced tense
b) in covered position single vowel letters are pronounced lax

98
belongs to not only 1 morpheme
99
belongs to only 1 morpheme
100
/a/ // /a/, they do not consist of a vowel sound plus a schwa //
41
II Spelling alternations:
1 Purpose of consonant doubling:
a) to retain covered position:
fit fitting, fitted; Jim Jimmy; Vic Vicky101
only after stressed vowels102: commt ~ commtted, forgt~ forgttable,
occr ~ occrring, but vmit ~ vmiting, trget ~ trgeting,
cnquer ~ cnquering
in BrE also after unstressed vowels for L: trvel ~ trvelling (not in AmE:
trveling);
optionally for S: bas ~ bas(s)ed, fcus ~ fcus(s)ing
covered position is retained even when
unjustified ( ecause the vowel is tense):

these changes are purely orthographical: the pronunciation of these forms is


perfectly regular, in fact, the spelling is modified exactly to reflect the fact
that the pronunciation does not change
2 Purpose of silent E:
a) to retain free position: hated, hating hate; even when this is unjustified
(because the vowel is lax): c ming ~ c me
b) to show that the previous consonant is soft: age, notice; if the vowel is lax, the
consonant is usually dou led103 b dge, dge, br dge, l dge, j dge (or not:
all ge, cf. allegation)
c) after word-final /v/:the letter V must be followed by a vowel letter: g ve, h ve,
l ve; accordingly any vowel followed by V is in free graphic position
III Quantity deviations:
1 Irregular quantity104 deviations 1 lax vowel in free position:
a) in the last syllable it is rare:

101
sometimes, C doubling is used in a way as Germans do (c ck)
102
the accents indicate the place of of stress
103
g is doubled as dg
104
don't have a lax/tense vowel when it would be expected

42
b) in the second-last syllable it is common and often unpredictable:
tom vs. d tum
d vil vs. vil
l mon vs. d mon
v ry vs. V ry
tr ple vs. disc ple
s rup vs. C rus
2 Laxing environments: except for U, vowels in free position are regularly lax in the
following environments:
a) third last syllable
b) in the syllable right before word-final -ic, -id, -it, -et, -el, -ishn,v
c) before free U
d) only I/Y before CiV105
3 Trisyllabic106 laxness:

105
Consonant + letter "i" + Vowel; usually in such an equation, lower case letters are real letters, do not
symbolyse anything, they are the letters (sounds) themselves.
106
3rd syllable

43
4 Laxing endings:

5 Free U laxness:

44
6 CiV laxness:

7 Prevocalic107 tenseness:

107
initial sound segment of vowels

45
8 CiV tenseness:

9 Overriding108 tenseness:

108
kiugr, kiemelked;

46
10 Irregular quantity deviations 2: tense vowel in covered position

47
The vowels of unstressed syllables
I Free variation109 and neutralization110:
1 Free variation exists in stressed syllables: either /a, i:/, apricot /eprkt,
prkt/
2 But it is much more common in unstressed syllables:
family /fmli, fmli:, fmI, fmI, fmi, fmli. . . / (the first,stressed syllable is
stable, the second and third unstressed syllables vary
3 Vowel contrasts are neutralized in unstressed syllables:
man men (// vs. /e/), but postman = postmen (// in both)
4 In an unstressed syllable: // can be spelt by any vowel letter
German, cypress, peril, martyr, abbot, fetus, villain, famous
II The fate of vowels111 in unstressed syllables:
1 In unstressed syllables:
a) vowels reduce: the reduced or weak vowels are / i j ju ju:/ (and perhaps
/o/; /i/ means // or /i:/, /ju/ means /j/ or /ju:/)
b) vowels may disappear leaving only their syllabicity behind: syllabic consonant
formation112 (C marks a syllabic consonant; C stands for consonant here, not part
of the sign)
c) vowels may disappear leaving only their place of articulation behind: high vowel
gliding113
d) vowels may disappear altogether: syncope114
2 /j/ may palatalize115 the preceding consonant:
NB: in the following environments pretonic means efore the stressed sylla le
III Single vowel letters in unstressed syllables:
1 A in an unstressed syllable:
a) is usually pronounced //116:
abut, ag, bnary, blative, mmbranous, bggar, lan, Fnland
b) is pronounced / / (i.e., // or //):
in the endings -ace, -ase, -ate, -ave: mnace, prchase, snate, ctave
in chracter
c) is pronounced // in the ending -age: vllage, mnage, hstage

109
replacing one language element by another in a particular unit does not change the meaning of the unit
110
vowel contrasts may disappear in particular environments, e.g. we have spill, still and skill, but not *sbill,
*sdill or *sgill, here, after /s/. (note the voiced-voiceless counterparts)
111
no matter how funny it sounds (:D), it means the end of Vs
112
it's a C which makes up another syllable without a vowel (or with a //), e.g. feel /fi:.l/, so 2 syllables, not
only 1
113
turns into a glide; in phonetics and phonology, a semivowel (or glide) is a sound, such as English /w/ or /j/,
that is phonetically similar to a vowel sound but functions as the syllable boundary rather than as the nucleus
of a syllable, cf: some pages further
114
unstressed syllables disappeare, cf: some pages further
115
cf: some pages further
116
Hungarian-like accents indicate the stress again

48
2 E in an unstressed syllable:
a) is usually pronounced / /:
erse, escpe, evnt, derve, dedce, serne, prblem, chcken, ctress, hpeless,
conctenate
b) is pronounced //nonpretonically before R, L: mderate, clery, pster, lvel,
philtely
c) is pronounced // in the last syllable (except -el, -er, -em, -en, -ess): hted, bses,
tcket,
d) is pronounced /i/:
prevocalically117: vdeo, petrleum, rea
word finally (if not mute): apstrophe, smile, s ncope
118
3 I/Y in an unstressed syllable:
a) is usually pronounced / /: fmily, cuncil, homnymy
b) is pronounced //: before a tautosyllabic119 R c nfirmtion120, Virgnia, tpir,
zphyr
c) is pronounced //:
word initially: ignte, illgal, innte, immne, Itlian
in the last syllable (except in -il(e)):
pnic, vlid, dlphin, dit, mking, rbbish, vctim, pposite, anlysis, ntice,
ngative
d) is pronounced /i/121:
word finally: bikni, khki, crry, crried, bby, bbies
prevocalically: opnion, ndia, hlcyon, hlium
4 O in an unstressed syllable:
a) it is usually pronounced //:
orginal, hstory, nchor, bllot, s nod
122
5 is pronounced /o/ :
a) word finally:
bngo, bro, jdo, mtto, Plto
cf. fllow, brrow, tomrrow
it is not clear if this is a reduced123 vowel

117
before a vowel
118
they have the same voice value
119
in the same syllable
120
is a secondary stress
121
difference etween // and /i/ is in length, /i/ is longer (well, in fact // is somewhat unlike the Hungarian
pronunciation of the letter "i", it's rather etween a HUNG "i" and a short "")
122
also can be transcribed as //
123
a weak vowel, like schwa // weak Vs: / i u /

49
6 U in an unstressed syllable:
a) is pronounced //:
in covered position124: bgus, grund, bluster, pndulum, buffon
before R: njury, zure125, perture, tnure
b) is pronounced /j / before consonant + reduced vowel:
stmulus, artculateADJ, cmmunist
c) is pronounced /j / or /ju/ before consonant+full vowel:
stmulate, artculatev, cmmunism
d) is pronounced /ju/ in prevocalic and/or pretonic position:
stution, rduous, stupdity, unte
e) is pronounced /ju:/ in the last syllable (exc. before R):
mdule, vlume, vlue (but as an exception, frtune has //)
IV Syllabic consonant formation: a C1C2 sequence may become a C1 2 sequence126
127
128

124
VC or VCC
125
NB! //!!! not /z/
126
so: [first consonant] + [sound ] + [second consonant] [first consonant] + [second syllabic consonant]
127
O struents: stops /p t d k g/, affricates / /, fricatives /f v s z h/
128
Sonorants: nasals /m n /, liquids /l r/, glides /j w)

50
V High vowel gliding:

129

VI Syncope130:
1 Syllabic consonants may lose their syllabicity:

129
palatality palate, la iovelar http://en.wikipedia.org/wiki/Labial%E2%80%93velar_consonant,
remains, MAY refer to the fact that the next syllable (usually a V) "starts" from the place of the glided sound (?)
130
hangugrats, like in Hung bok(o)rok

51
2 Some cases where the syncopated form is lexicalized:

3 In fast speech: syncope occurs also in


a) pretonic position
b) before an obstruent, and in other cases where the second C could not have been
syllabic

(crossed-out sounds are NOT pronounced)


VII Palatalization:

NB: the dagger-like symbol (bc it's called a dagger) indicates that the form is obsolete,
presumably nobody speaks that way (maybe elderly ppl only)
VIII Mergers131:

131
loss of distinction between phonemes

52
Stress & intonation
I Suprasegmental properties of speech:
1 Segmental and suprasegmental properties:
a) segmental properties: these properties of speech can be associated with
individual segments (i.e., sound or phonemes): e.g., voicelessness belongs to a
given obstruent, backness to a given vowel
b) suprasegmental properties: these properties are associated with longer strings of
segments (e.g., a syllable)
II Stress:
1 Phonetically132: stress is an increase in the loudness (or intensity) of the speech
signal133: a rise in the amplitude of the sound waves
2 Phonologically: stress is a relational (not an absolute, rather it's relative)
phenomenon a given portion of the speech signal is louder than another, i.e., it is
more stressed. This means that one cant tell of an isolated portion of the speech
signal whether it is stressed or not, this can be decided only in comparison with
other portions
3 The carrier of stress: the common wisdom is that stress is a property of syllables, but
there is also reason to assume that it is vowels that are stressed
4 Degrees of stress: phonologically four degrees of stress are relevant in English

ictus: it's like a "hit on the table while speaking"; a stress in this respect, primary and
secondary stress is identical (stress given by the rhythm, ritmus adta hangsly)
a) the difference between primary and secondary stress is predictable:
the last ictus is primary stressed
all previous ictuses are secondary stressed

132
Note that while phonetics deals with production, physical properties, perception of the actual sounds and
the suprasegmental elements of speech; phonology deals with is the study of the phonemes and their
combinations in words and morphemes, and also of the discrete suprasegmental elements in words and
sentences. They are not the same, only somewhat related to each other.
133
the actual sounds of speech which can be heard

53
5 Marking stress: 4 ways to do this

syllabic boundaries within the feet134:


e -lectroen - cepha - lography

a) foot:
consists of a major stressed syllable (head of the foot) and the following non-
major-stressed syllables
there are typically two syllables in a foot (trochee /x)135 or three (dactyl
/xx) feet with more than three syllables are rare
initially a degenerate foot (lacking a head, like the first syllable in the
example) may occur
6 Properties of stress:
a) no major stress after primary: this is true by definition the last major stress is
primary stress
b) a major stress must fall on one of the first two syllables of a word: EARLY STRESS
REQUIREMENT: *#00, *#03, *#30, *#33136 no English word starts with 2
unstressed (no - ictus) syllables
c) content words: (a.k.a. lexical words, i.e., nouns, verbs, adjectives, adverbs) must
contain at least one non-zero-stressed syllable
d) major stress on adjacent syllables is dispreferred: within a word, 321 stress
pattern does not exist, within a phrase/sentence, 321 stress pattern becomes
231, etc. (i.e., feet contain more than a stressed syllable if possible); more
generally stress is preferably as evenly (egyenletesen) distributed as possible,
cases where it is not evenly distributed are referred to as STRESS CLASH137

134
the next foot starts at the next stressed syllable, so cat in a hat is 2 syllables (cat in a and hat), cat flap is
also 2
135
note that / stands for stressed, while x stands for unstressed syllables
136
these mean the level of stress for each syllable and the * indicates that they are all ill-formed modes; a 2 or
a 1 MUST fall on one of the places in #_ _ position
137
A stress clash can occur when two stressed syllables are too close to each other. For example, the word
'nineteen' spoken in isolation has stress on the second syllable. But when it is placed before 'girls' the stress on
'nineteen' can shift to the first syllable.

54
7 Stress adjustment in phrases:

(will be no stress, where there haven't been any before the shift)
8 Locating stress in the word:
a) primary stress: primary stress usually falls on one of the last three syllables
b) place of stress: the place of stress depends on the weight of syllables (the length
of their vowel and the presence of a
syllable final consonant)
c) the place of stress usually differs in nouns and verbs of the same shape 138:
e.g., trmentV vs. tormntN
d) there are very many exceptions to the stress-locating algorithms, therefore some
analysts consider stress lexically specified (so each stress pattern should be
learnt with the words itself)
e) the basic rule: the stressed syllables of content words are stressed, that of
function words (e.g. will, can, would etc.) are unstressed
f) even distribution of stress is achieved by:
STRESS DELETION: the ictus is deleted from a content word, e.g.,
S e k ssed Je! S e k ssed Je (note that content words may lose their
ictus (major stress), but must remain tertiary (minor) stressed)
STRESS CREATION: a function word is assigned an ictus, e.g.,
She has to retrn! She h s to retrn, cf. S e has to retrn, where has
remains unstressed

138
this is called an approximate conversion lexemes undergo a small but systematic change in pronunciation
and are thereby assigned to a different word class. Sometimes this small change is a stress shift.

55
g) weak forms: since monosyllabic139 function words (prepositions, conjunctions140,
auxiliaries, determiners, pronouns) are usually unstressed most of them have a
weak (i.e., unstressed) form:

h) monosyllabic function words that do not have a weak form include: off, on, my, I,
you, they, may, might, ought
III Intonation:
1 Phonetically: intonation is rendered by pitch change, i.e., rising or lowering of the
frequency of the speech signal
2 Phonologically: tones distinguish
a) modalities, attitudes, (non)completeness in intonation languages (like English or
Hungarian)
b) lexical words141 in tone languages (like Chinese or Yoruba)
3 The carrier of intonation:
a) the last part (the nucleus) of the tone unit,
b) the primary-stressed syllable of the phrase or sentence
4 Parts of the tone unit:

1. the pre-head: a degenerate foot before the first stressed syllable of the tone-unit
2. the head: the first stressed syllable of the tone-unit
3. the body: all syllables from the first to the last stressed syllable not including
either
4. the tonic: (nucleus/sentence-stress), the last stressed syllable of the tone-unit,
characterized by pitch change
5. the tail: all unstressed syllables after the tonic

139
consists of only 1 syllable
140
e.g. and or but
141
in such languages, if an ortographically identical word has 2 intonations, they are 2 individual words

56
5 Types of nuclear tone: `falling, ' stress, rising, falling-rising

57
Phonotactic constraints142
I Element combinations: given a set of words some combinations are possible, others are not e.g.,
of {eats, nuts, Tom} we get:
1 Correct:Tom eats nuts, nuts Tom eats
2 Incorrect: (the * stands for incorrectness in linguistics) *Tom nuts eats, *nuts eats Tom *eats
Tom nuts, *eats nuts Tom
3 In phonology: a similar sensitivity to order is observable in phonology, e.g., of {e, n, t}, we
get:
a) correct: ten, net, ent (a mythological being)
b) incorrect: *tne, *nte, *etn
II Possible/impossible words, existing/nonexistent words:

1 Phonotactic constraints: They are meant to filter out impossible words (actually,
mission failed bc we have schwa...:D)
2 Accidental gaps: many words not filtered out by phonotactic constraints still do not
exist143
a) some words are filtered out because they contain a rare string, still they do exist
(e.g. schwa), while others (e.g. bnick) do not (as /bn/ is considered to be a rare
string in english)
III Phonotactic constraints:
2 Within morphemes: they help to decide whether it exists or not
3 Across morphemes: sound combinations occur much more freely, so /kb/ does not
occur within a morpheme, but it does in blackboard, where a morpheme boundary
separates the cluster (so it gives no clue)
4 Sound organization is systematic: phonotactic constraints are normally formulated
with reference to natural classes144

142
knyszersg, knyszerts
143
actually, blick is an existing word, as Eric Bakovic notes:
http://camba.ucsd.edu/phonoloblog/index.php/2007/04/05/did-you-mean-brick-block-brick (note my note :D)
144
In phonology, a natural class is a set of sounds in a language that share certain phonetic features. The sound
system of every language includes several natural classes, each distinguished from other classes by certain
features. A given natural class is described using the minimum number of features needed to include all sounds
within the class and exclude all sounds outside the class. For example, the set containing the sounds /p/, /t/,
and /k/ is a natural class in English, namely voiceless stops. This class contrasts with several other classes, such
as the voiced stops or voiceless fricatives etc.

58
a) Typically occurs:
at the edges of words
between two consonants
between two vowels
between a vowel and the following consonant
BUT: rarely holds between a consonant and the following vowel, usually only
adjacent (=neighbouring) segments are affected
2 Word edges:
a) the beginning of the word:
vowels: // and /u:/ are rare (umlaut; oodles, ooze) but this does not seem to
be systematic (rarely occures)
consonants:
// does not occur
#CC is: (2 consonants in the beginning of the word)
o /s/+{p t k m n l j w}
o /r/
o {p t k b d g f T m n h}+/j/
o {p t k d g f }+{l r j w} (nonstrident obstruent145+nonnasal sonorant),
but the following do not occur {p b f}+/w/, {t d }+/l/
(a) #C1C2C3 146is possible only if:
C1=/s/ and #C2C3 is also possible (straytray, splayplay), but not vice
versa: #twin*#stw
b) the end of the word:
vowels: plain lax vowels do not occur a word-final stressed vowel is long
(broad or tense)
consonants:
/r j w h/ do not occur
CC# is:
o homorganic147 nasal + obstruent: /mp nt n k *mb nd n *g mf
*mv n *n ns nz/ (/n n/ rarely occurs)
o /lp lt l lk lb ld l *lg lf l ls lv *l lz lm/ (/l ln/ rarely occurs)
o fricative + stop: /sp st sk ft/
o stop + stop: /pt kt/
o stop + fricative: /ps ks/ (no voiced obstruent + obstruent clusters)
CCC# is rare: /mpt kt lpt lkt mps ks/ may occur

145
Strident sounds are produced with a complex constriction (sszehzds, sszeszoruls) forcing the
airstream to strike two surfaces, producing high-intensity fricative noise (rshang); nonstrident sounds are
produced without such a constriction. (Sibilants, labiodentals, uvulars vs. all other sounds.)
146
3 different consonants in the beginning of a word
147
having the same place of articulation

59
3 Sonorants: (crossed-out letters represent unuttered sounds in SSBE)
a) /r w j/: only occur before a vowel or sylla ic consonant rack, car engine

b) /h/: only occurs before a stressed vowel or at the beginning of a word: horzon,
lcoh l, vehcular,
c) //: only occurs before {k g} and word-finally: ink /Ik/, finger /fIg/, fang /f/
4 Vowel clusters:
a) diphthongs: the diphthong inventory defines possible vowel + vowel sequences:
e.g., /i * e *e a a I o * u/ (NB: /i/=/i:/, /u/=/u:/)
b) heterosyllabic148 vowel clusters: the first vowel must be plain tense (or
unstressed /i u/), prevocalic tenseness149: *V, *V, *eV, *V, *V, *V, *V
5 Some vowel + consonant restrictions:
a) tense vowels do not occur before consonant clusters if either of the consonants is
noncoronal150: * mp, * ft, * lk, etc.
b) tense vowels, broad vowels and schwa () do not occur efore // (the velar
nasal () behaves like a noncoronal cluster)
c) only coronals occur after /a/ flout, loud, couch, gouge, house, arouse, south,
Louth, brown, owl, but *am, *ap, *ak, etc.
d) only alveolars occur after /I/ exploit, void, voice, noise, coin, coil, but *,
*, *, *p (exc. coif)
6 CC+V restrictions: while consonant + vowel restrictions are rare, the following occur
in English:
a) Consonant + j can only be followed by /u: : /: cf. pubic, cure, tutorial,
million
b) Consonant + w cannot be followed by /u: /: cf. two /tu:/ (the /w/ drops, but it
is present in twelve, twenty, twice, between, etc.)
NOTE: C+V restrictions are NOT presented here, but they exist
IV Language change: one aspect of language change is the modification of phonotactic
constraints, e.g.,

148
it belongs to not only one syllable
149
see lcture note on graphic positions
150
Coronal consonants are consonants articulated with the flexible front part of the tongue. Only the coronal
consonants can be divided into apical (using the tip of the tongue), laminal (using the blade of the tongue),
domed (with the tongue bunched up), or subapical (using the underside of the tongue), as well as a few rarer
orientations, because only the front of the tongue has such dexterity. (see more at:
http://en.wikipedia.org/wiki/Coronal_consonant)

60
V Loanwords:
1 Alteration: loanwords are usually altered to fit the phonotactic constraints of the
host language

2 Modifying role: but loans may modify phonotactic constraints

61
Accents of English
I Some facts about dialects: there is no linguistic value difference between dialects
1 Role of prestige: the prestige of standard dialects is sociological/historical
linguistically no dialect is clearer, purer, or more logical than any other, there is no
criterion to select the right form of latter from among:

(flap/tap151, 152, glottal stop153, /e/ assimiliated to /r/)


2 The difference between language and dialect: It is unclear; mutual intelligibility could
be a criterion, but it is not a transitive relation: if A B & B C A C speakers of
dialect A understand speakers of dialect B and speakers of dialect B understand
speakers of dialect C, yet speakers of dialect A do not understand speakers of dialect
C (and also: note that ppl from Denmark and Sweden mutually understand eash
other)
a common wisdom: "a language is a dialect with an army and a navy"
3 Why do people speak different accents? to project particular images of themselves
language use is just like clothing, hair style, or any other means of communicating
ones identity
4 Criticising other peoples language use: It is called linguicism; linguists see this as
improper behaviour, but it is a common prejudice in many societies (including
Hungarian154):
aki azt mondja, hogy eszek-iszok, az egy kznsges piszok
aki suk-skl az egy gaztettekre is kpes
unfortunately, such ideas are still to be found even in educated circles
II Language variation:
1 Languages usually vary:
a) by geographic location (this is becoming less so as people move around and
communicate more freely)
b) by social status
c) by age, sex, profession, etc.
d) in time (variation by age partly reflects this fact: younger people speak differently
than older people, older people die first, so the language changes)
in fact, no two people speak language in exactly the same way (everybody speaks

151
In most US accents the letter <t> is pronounced as an r-like sound, so e.g. city as /sii/
152
An allophone of the /r/, it is like the diphthong // but accompanied with lip rounding. (???)
153
In some GB accents (and of course, in some others) the letter <t> is omitted and a gap is left behind realised
as a glottal stop, e.g. latter as /l./(the comes instead of the ".")
154
For exchange students: these are regarded as nonstandard in HUNG associated with "lower social strata"
st
and different ethnic groups; in the 1 S, the phrase "eszek-iszok" is in nominative case instead of the standard
nd
"eszem-iszom" in accusative case, while in the 2 one, the term "suk-skl" denotes the phenomena when
HUNG natives speakers use verbs in imperative istead of the standard declarative in statements.

62
their own idiolect155), but linguists are interested in systems, not individual
incidences.
III Accent vs. dialect:
1 Dialect: variation involving morphology, syntax, lexical units
2 Accent: variation involving only pronunciation (accent also means stress, or the
diacritics used to mark stress, , etc.)
3 Accentual differences:
a) Standard Southern British English (SSBE):
You must eat it up sounds as /ju mst 'i:t t 'p/
b) only accent is different:
Standard Northern British English (SNBE): /j mst 'i:t t 'p/
West Indies: /j m st 'i:t t ' p/
4 Dialect and accent are both different:
a) traditional NBE: /a m 'gr t ':tn/ (You must eat is up hardly understood)
b) Jamaican Creole: /j m s 'njam 'a:f/ (You must eat it up)
5 Types of accent differences:
a) irregular: involving individual lexical items
SSBE GA156
SSBE // is generally GA / :/, but in
what it is GA //

b) regular: involving the majority of items in its scope


systemic
realizational
distributional
6 Realizational accent differences: a realizational difference between accents does not
involve the phoneme systems of the two accents, they differ only in the actual
pronuncation using the same phonems e.g.,
a) SSBE /e/ and /o/ is Scottish Standard English /e/ and /o/: gate SSE /get/, goat
SSE /got/
b) SSBE // is New Zeland English (NZE) //
c) SSBE /p t k b d g/ is Merseyside English / s x z / (between sonorants)
in themselves, these changes do not change phoneme inventories the system

155
The total of all the varieties of a language that a person knows is the personsidiolect.
156
General American, the most widely spoken variety of ENG in the US

63
remains unchanged (cf. Saussures chess analogy: the rook (bstya) may be
replaced by a piece of rubber, the game does not change)
7 Systemic accent differences: a systemic difference merges contrasts present in other
dialects and retain others merged in other dialects. A phoneme opposition is present
in a variety but absent from another.

Saussure: if the rook and the knight become


indistinguishable, the rules of the game
changes

8 Distributional accent differences:

64
IV Nonnative accents:
As the de facto international language, English has the greatest number of nonnative
speakers (speakers whose first language is other than English) in the world; accordingly,
English has many nonnative accents, among them Hunglish
There are other similar dialects/accents like Engrish (Far-Eastern variants), Finglish
(Finnish)
The term Hunglish and its counterparts are often used to denote a dialect (not only an
accent), and they are often used pejoratively (negatively), however, (as Professor
Ndasdy writes) it is still better to have a systematic Hunglish accent than an ill-
prounced English-sounding one, as natives more easily unerstand it.
1 Hunglish: Hunglish is useful for an effective Hungarianization of English loanwords
it replaces sounds not occurring in (standard) Hungarian by others which do, e.g.,

lD prjMErili inTENded for inKORprting INglis


LNwrdz nd NMZ in hanGErin TEKSTS, disz KJND v
prNANszisn kn iFEKtivli JZD tu kMJnikt vid
INglis SZPkrz. di EKsznt iz KLRli Ntisz li hanGErin,
bat kSZEPt l t d szm TJM. Lsz, der iz n ND tu
MEmrjz l dz FENszi Jpj SZIM lz!
(capitals indicate stress)

For those how are not familiar with the HUN ortography:
Although, primarily intended for incorporating English lownwords and names in Hungarian
texts, this kind of pronuciation can effectively be used to communicate with English
speakers. The accent is clearly noticeably Hungarian, but acceptable at the time. Also, there
is no need to memorise all those fancy IPA symbols! :D

65
66

You might also like