Midterm Review
Introduction to Phonetics, L306
R. Port, March 5, 2008
Text Materials:
- Ladefoged 5th Edition,
Chapts 1-7 (and
handouts).
- American English
allophones webpage
- Dynamics of Speech handout
- ToBi materials
Midterm Exam on Friday: Mostly
objective questions (fill-ins, short answer), some diagram labelling, a
`dynamics of speech' diagram or two. Probably a Tobi phrase or
two and a short transcription. (So bring your headphones)
Transcription: Be able to
transcribe dialects of English using the
basic IPA alphabet. Recognize implosives, ejectives, clicks
(labial, lateral and palatal when voiced, voiceless, aspirated and
nasal).
Speech anatomy: lips, tongue tip
(apex), tongue
blade, tongue dorsum, tongue root. Mouth, nasal cavity, pharynx
(nasal
and oral), soft palate (velum), hard palate, alveolar ridge.
Larynx vs.
glottis (larynx is cartilage and muscle, the glottis is a space),
vocal folds, false vocal folds, trachea, epiglottis.
Basic symbol set of the IPA (International
Phonetic
Assoc.) alphabet.
Places of Articulation (for
consonants) - at two
`levels of detail':
Labial, Apical, Palatal, Velar, Glottal, OR bilabial,
labiodental,
dental, alveolar, retroflex, palato-alveolar, palatal, velar, uvular,
pharyngeal,
labiovelar, glottal (Ladef, Table on p. 167)
Manners of Articulation (for
consonants) (Chapters
3, 7, see `Manner' Table
on p. 176): stop, nasal, approximant, fricative, trill, tap, flap,
lateral
vs. central. What argument could you raise to claim that the
click property is
a Manner of Artic? (because they occur at different places, eg, labial,
dental, palatal, etc) What argument could you raise to claim that
the
click property is a Place of Artic? (because they come in different
manners, like nasal, +/- voice, ejective, lateral, etc).
But the Place
x Manner
x Voicing
model of phonetics is inadequate
because:
1) `secondary articulations'?, clicks? (This
scheme does not allow enough independent dimensions)
2) empty cells: eg, `lateral velars,
velar flaps'? (The
dimensions
of Place and Manner are often not independent)
3) No place for timing patterns like VOT
or
C length. (`Nonsegment-like
properties' are ignored)
Vowels
Vowel dimensions: tongue height, tongue
backness, lip
rounding. For English, how many V dimensions? Only 2? (maybe just
height and backness) Or 3? (height, backness, rounding)
Or more? (height, backness, rounding,
length, etc).
Monophthongs vs diphthongs. Which Vs in
English
are ambiguous on this property?
`Secondary vowel articulations': nasalization,
rhoticism
(or retroflexion)
Stressed, unstressed; reduced vowels, full vowel;
pitch
accent.
Tense Vs, lax Vs; closed vs. open syllables (Which
Eng
Vs only occur in closed syllables? Which contrasts are usually
neutralized before /r/?
Which English Vs are the
longest in duration and which the shortest?)
Consonants
Stops, fricatives (obstruents); Essential
gestures: oral
closure + nasal closure OR glottal closure
Apical stop variants: lateral release (bottle), glottal stop (cotton), flap
(or
tap) (butter), trill,
velar release (sudden)
Homorganic relationships, voicing pairs. What is
coarticulation?
Air stream mechanisms: (pulmonic, glottalic,
velaric)
X
(ingressive, egressive). Match these with `plosive, ejective,
implosive, click.'
Voice-onset time: Continuous-valued, in
fact, but
across
languages has 3 modes: prevoiced, short-lag (unaspirated), long-lag
(aspirated).
Cf. Eng vs. French VOT
patterns.
Other `voicing' cues for English: vowel
duration
and obstruent duration for postvocalic obstruents.
Affricates vs. fricatives (note spelling!)
Breathy-voiced stops (voiced aspirates) (eg, in
what
languages?)
Glottal articulations: [h] vs. voiceless vs.
voiced
(laryngealized voice, glottal fry, breathy voice, etc.)
Be familiar with the major English `allophonic variants' of vowels and
consonants (pp. 71-76).
Port's `Continuous-Time English Speech
Production Model'
Based on a set of `independent
articulatory systems': Lips, Tongue-tip, Tongue-body, Velum,
Vocal-folds. Each of these
dimensions
(or systems) is largely independent of the others, and has a limited
set
of mutually exclusive values. (Eg, the lips can be open, rounded
or
closed, but can never be in two states at once.) Each system changes
state
in continuous time in one dimension
(eg, lip rounding) or two
(eg, tongue body
position). Plots of each articulatory
system (reducing each to a single dimension) can be plotted against
time
(somewhat like a musical score). Be able to draw such plots for any
single
words
showing coarticulation, assimilation and certain prominent timing
details.
Properties of the IPA phonetic alphabet
- IPA attempts to be universal,
but is
really (according to Port) only more
consistent than orthography (that
is, than conventional spelling)
- Good for writing down approximate
actual pronunciations (if you were trained to use an alphabet)
but not exact ones.
- May be a better
model for cognitive
form of words
than orthography, but is not necessarily an adequate model.
- It's feasible to cover most
of the sounds
of languages
of the world, but impossible to get all of them (claims
Port) because there is no reason to believe there is a fixed inventory.
(Ie, languages can keep inventing new sounds, just like new tunes on a
fiddle.)
- The closer we look, the more subtle
features
we find that
differ between languages (The phonetic space cannot be
closed
but
rather must be open-ended) (Phoneticians would generally
agree
with this but not most
linguists.)
Phonology: A. the application of
phonetic
sounds for spelling
morphemes in languages.
B. study of the sound patterns
employed in the dictionary and speech of a language community
(Port's def)
- People seem to have awareness
of gross
sound categories,
like phonemes, but not to be aware of phonetic details. (This is
probably because of literacy training.) Though some
details
are easier to hear than others. Eg, its easy to hear stop [t] vs.
flap, or longer V in bad than bat. But its hard to hear
/k/
longer than /g/ in backer-bagger. Speakers of
language A will often find some sounds in language B very difficult to
impossible to differentiate or identify - thus there are NO perfect
transcribers.
- Phonemes:
hypothesized abstract,
cognitive sound units
that strongly resemble the letters of an alphabet. Units like
these are
thought to
``spell'' words in human memory. They are nonoverlapping, serially ordered, always either identical or distinct. And
each word supposedly
has a single citation-form spelling.
(But I think these are really
properties of our alphabetical notation. These are
obviously not properties of
speech sound. Linguists (and many phoneticians) prefer to ignore this
fact.)
- Prosody: Languages
have conventional
patterns of pitch,
loudness and timing at the level of words and phrases. But these are
typically
difficult to describe for most languages - given current knowledge. One
problem is that they cannot be reduced (very well) to letters - though
ToBI tries to approximate this.
- Allophonic rules
sometimes cause `alternations' in certain words (eg,
Ladef, p. 39). By speaking of allophones, we are able to spell
vocabulary using a smaller (and supposedly more efficient) set of
symbols. (Do speaker-hearers store words in this `efficient
form? Almost certainly not -
says Port.) Some allophonic rules
are language specific (eg, Eng nonaspiration of /t/
after /s/ in store), some nearly universal (eg, nasalization of
V/_N, undershooting a low vowel between high vowels, eg, in yaya vs. baba - lower [a] in baba). Be familiar with the major
English allophonic rules in Ladefoged pp 71-76. Be able to
interpret the rule notation used by Ladefoged.
- Systematic phonetic transcription vs. Impressionistic phonetic
transcription. Assumes a phonological analysis. Systematic
implies using the allophonic rules in the transcription,
Impressionistic implies just listening and recording what you actually
hear.
- Feature analysis of phonemes:
know the
main phonetic
features of the consonant phonemes of English (although not the binary
ones used
by phonologists).
Prosody (other than timing)
ToBi
Analysis (Tones and Break Indices). This is the most useful
method
for transcribing English prosody (and some other languages as
well).
There are 4 types of tone: Pitch accents (PA)
(H* peak,
L*
low, L*+H scoop, L+H* rising peak, (H)...!H
downstep high), Phrasal tones (H-, L-),
final boundary tones (L%, H%) and initial boundary
tones (%H, %L). Some common
patterns are the ``Declarative phrase intonation'' (%L..PA..L-L%), List item or continuation rise (..PA..L-H%)
and ``Nuclear accent'' (a pitch accent near the end of an intonation
phrase).
RFP