Review for Phonetics and Phonology
Caveats:
1) You must not confuse letters (graphemes) with sounds (phones).
2) You must not allow the orthographic representation of a word (i.e., spelling) to
interfere with your perception of the actual sounds you make when you say the word.
You do not have to account for the spelling of a word—don’t try to justify the presence
of a g in the word night. That’s not your job in phonetics and phonology. (It would only
be your job if you studied the historical orthographic sound/symbol correspondence of a
language).
3) Remember that / / are for phonemes, [ ] are for phones/allophones.
4) Remember that the carat/^/ is stressed and the schwa / / is unstressed.
5) Basic consonant feature bundles: [+/--VOICE], [PLACE], [MANNER]
6) Basic Vowel feature bundles: [HIGH, +/--BACK, +/-- , +/-- ROUND, TENSE/LAX]
Broad Transcription: phonemic level
Narrow Transcription: phonetic level—
1) what specific allophones of a phoneme are we looking at.
2) use diacritics to indicate the allophonic features of the segment
Monophthongs:
Remember that the carat /^ /is a real vowel that occurs in stressed syllables:
luck, mother, rough, cut, chug, stuff, shove, crud, puppy, Elmer Fudd, pun
R-Coloring
All vowels get changed a little bit when they are next to an /r/ or an /l/.
You can’t really keep a pure /o/ when it is colored by an /r/-- so we use open /o/ with /r/.
snore, pour, boar, glory, poor
For a carat +/r/ which occurs in stressed syllables (or you might think of it as an /u/ + R)
in words like bird, curd, fur, earth, sure, cure, learn we use the symbol that looks like a
3 with a small carat-wing on the top right
For a schwa + /r/ (which only occurs in unstressed syllables) in words like father, later,
perform, you use the symbol that looks like the schwa with the same little small carat-
wing on the top right
Use pervert as a memory jogger.
Diphthongs:
We are going to recognize three diphthongs in English : /aI/, /au/, and /oI/.
Try a cowboy.
(Some books will recognize the long /ou/ in Boat or the long /eI/ in Bait; we won’t.)
In phonetics you’re describing a sound physically. You give its actual DNA, size,
physical characteristics. In phonology you assign it a race or an ethnicity. No two
phones are the same, but you can classify them according to common features or
characteristics.
In phonology, you decide which of a finite number of boxes to put that phone (sound) in.
You assign it to a matching phoneme. Sometimes a phone may look like one phoneme
but be classified under a different phoneme such as the /t/ in butter that sounds like a /d/.
We still ‘hear’ it as a /t/, even though in reality it’s obviously not the same sound as the
initial phone in top, or the second phone in stop. The [d] sounding /t/ then is an
allophone of /t/. It’s the phonetic realization of the phoneme /t/ in a specific phonetic
environment—in this case, between two vowels.
Phone: a physical acoustic reality
Phoneme: a psychological reality, the category or ‘letter’ speakers of a language perceive
a sound to represent.
Allophone: the phonetic realization of a phoneme which might be different from the
unmarked prototypical realization of that phoneme. For example, /t/ doesn’t sound like a
/t/ in butter, cotton, or baton as it does in stop.
Feature Bundle: the set of descriptors that identify a given phone
Orthography: System of written representation of a spoken language (and/or signed)
Grapheme:
The Feudal Stressed Vowel and its Schwa The Social Democracy of Vowels
Fiefdom
Rhythm timed sound systems: Centrifugal Syllable timed sound systems: Centripetal
Marked (less common among the world’s Unmarked (most common among the
languages) world’s languages)
English, German, Russian, French Japanese, Swahili, Spanish
Might Makes Right: Equal Opportunity Vowel Accommodation:
Well endowed vowels (those that occur in Every vowel gets equal time
stressed syllables) get even more than their
full value (1.3?), while their unlucky fief
vowels (those occupying the surrounding
lexical territory) get reduced (i.e., squashed
out of their value) into the schwa!
Seem to have deeper orthographies. . . ? Seems to be more commonly represented by
shallower orthographies
Remember that like just about everything else, we’re talking about a continuum here—
Yes Spanish is more syllable timed than English, but it’s a little less so than Japanese.
When you pronounce my first name (Elisa) in English, it starts and ends with a schwa
because the middle syllable [i] gets the primary stress. When you pronounce it in
Japanese, each syllable gets just about equal time and the first and last syllables are full,
the lax front mid vowel [E], and the low back [a], respectively. In Spanish, the first
syllable is the same as in Japanese, but the last syllable is a little closer to a schwa than
[a], in contrast to the Japanese pronunciation which is a purer [a].
In most languages a syllable consists of a Consonant and a Vowel. We describe them as
CVCV languages. In rhythm timed languages there can be clusters of C’s without V’s.
That results in syllables that get more press than others—syllables that get sounded
longer than others.
A syllable consists of an ONSET (C) a PEAK (V) and a CODA (C), where the onset and
coda are optional, but the peak is not optional. In other words, every normal syllable in
the world’s languages must have a vowel, whether or not it has any consonants.
In English the main exceptions to this are certain instances of /l/ /m//n/, and /r/ which are
all extra sonorant among consonants. These can all become syllabic which means that
they have enough voicing to carry a syllable by themselves without any schwa support
(although in the broadest transcription you can represent these syllables with a schwa in
front of the respective consonant). Syllabic consonants belong to the detail of narrow
transcription (where you identify the specific allophone of a phoneme instead of just put
the phonemes).
/l/ bottle, turtle, bubble, cuddle, purple, buckle, beagle, bugle
/m/ prism, spasm
/n/ bitten, prison, cotton, Pippin, , mission, vision
/r/ shopper, slobber, bicker, bigger, bidder, butter,
Some of the differences between two phones are random, or meaningless. We call that
Free Variation. If we say that a variety of a sound is an instance of free variation, we
mean that it doesn’t make a difference in the meaning of the word—it doesn’t make the
word into another word. For example, if you sometimes say envelope with the first
syllable sounding like the preposition ‘in’ and at other times pronounce it like the
preposition ‘on,’ it doesn’t confuse anyone about the meaning of the word. There are not
two different words based on that difference of pronunciation. You can also say spit and
sphit, but since [p] and [ph] are not two separate phonemes in English and since the
aspirated p never occurs except word initially, an anomalously aspirated [p] will just be
normalized and ‘heard’ as a [p].
On the other hand, when two phones cannot be interchanged without altering the meaning
of the word (i.e., making a different word), then you know you have two different
phonemes. If you are studying another language and you want to determine whether two
phones are separate phonemes or allophones of the same phoneme, the first and easiest
test is to look for minimal pairs.
Minimal pairs are two words that sound exactly the same except for one sound segment
(when they occur word initially they rhyme). For example, bat and pat are minimal pairs
in English. They are two different words with different meanings, so we know that /b/
and /p/ are separate, distinctive phonemes in English. On the other hand, pat and phat are
not two separate words in English, so we know that [p] and [ph] are not separate
phonemes in English, but are random variations (free variation) of the one phoneme /p/.
(although they very well may be in Thai, where they are two separate phonemes)
Caveat: the fact that there are homophones (the opposite of minimal pairs) in the
language, whether spelled the same or differently, doesn’t have any ramifications for the
sound system of that language. [Consider this: you walk into a room and see a mouse.
You are sure there is a mouse in the room. This is like finding a minimal pair and being
sure that you have two phonemes. Alternatively, you walk into a room and do not see a
mouse. This does NOT ASSURE YOU that there is not a mouse in the room, it just
means you are going to have to conduct a search before you can be sure.
Superman/Clark Kent exist in Complementary Distribution. They cannot exist in the
same place at the same time. Clark Kent is the phoneme, the unmarked form of the
phoneme which occurs most commonly and in the most environments. Under a narrow
set of exceptional circumstances (i.e., someone is in danger and need’s a superhero’s
services, which happens to me all the time) Clark Kent (entering the PHONE BOOTH,
hardy har har) takes the form of Superman. Same dude, different duds (and a slightly
different persona). Basically Superman is the allophone because he is rare and shows up
in only circumstances that are life and death situations. His appearance is therefore
PREDICTABLE. In formal phonological symbols we could write this:
/Clark Kent/ > [Superman] / [Death threat] ___ [Happy Ending]
Verbally: Clark Kent turns into Superman in the environment of: just after a death threat
and just before the happy ending.
The different forms of water is another metaphor for complementary distribution: steam,
water, ice—they are all different forms of the same substance, H2O, like being allophones
of the same phoneme whose form will be predictable, determined by their environment.
In the case of H2O, the environment that determines its form is the temperature. In the
case of a phoneme, the environment that determines its allophone is the class of sounds
before and/or after the slot for the target phone.
The opposite of complementary distribution is overlapping distribution. That means
two phones are treated as totally separate phonemes by the language under investigation.
To appeal to our Marvelous Metaphor: Clark Kent and Lois Lane are perfectly capable
and comfortable occurring in the same place at the same time. They can be a minimal
pair. The identification of their minimal pairiness assures us without further investigation
that they are in fact two totally different people; we don’t need to check their DNA or
anything tiresome like that.
Minimal pairs two words that have the same phones except for one segment (in the same
part of the word). The existence of a minimal pair proves that the two differing phones
belong to totally separate phonemes. Not having a minimal pair however, doesn’t
necessarily prove anything. (If this doesn’t make sense, think about this: walking into a
room and seeing a mouse confirms that there is a mouse in the room. Walking into a
room and not seeing a mouse doesn’t mean there is no mouse present, you just have to
investigate further).
word initially word medially word finally
nap and map snip and snap, mat and man
/n/ & /m/ /I/ and /ae/ /t/ and /n/
clap and crap
/l/ and /r/
What’s the formula for writing a phonological rule?
What are the abstract symbols and what do they mean?
The phoneme is considered to be the allophone that is unmarked (most common). The
marked allophones are called allophones of the phoneme.
So a phone is a generic uncategorized phone. When you analyze its usage in a given
language’s sound system you can determine whether it is a phoneme, an allophone of a
phoneme, or a random phone that is not recognized by that language (in which case it
will either get ‘normalized’ into an existing phone booth (treated as an allophone of a
recognized phoneme) or rejected altogether—the hearer will not recognize the word in
which it appears as a real word in her language.