Slide Sound structure Part II: Phonology

Sound Structure Sound Structure Part II Phonology 2009 Review of Phonetics Speech sounds are decomposable into articulatory primitives (also known as features) Consonants and Vowels Feature differences (e g , voiced vs voiceless, nasal vs not nasal, labial vs alveolar vs velar) lead to the diversity of sounds across languages We seem to perceive speech sounds as discrete units rather than continuous acoustic signals Seeing Speech Further aspects of sound Cognitive How important is speech to lang.

Sound Structure

Part II: Phonology

2009

Review of Phonetics

• Speech sounds are decomposable into articulatory

primitives (also known as features)

• Consonants and Vowels

• Feature differences (e.g., voiced vs voiceless, nasal vs

not nasal, labial vs alveolar vs velar) lead to the diversity

of sounds across languages

• We seem to perceive speech sounds as discrete units rather

than continuous acoustic signals

Seeing Speech

Further aspects of sound:

Cognitive

• How important is speech to language?

• The phoneme: the basic, distinctive sounds of

a language

– What it means to be distinctive

• How words are organized into subparts

(syllables and other units) for the

determination of stress

• Knowledge of language the internal

“grammar” in your mind will get more and more abstract than the physiological aspects

of language

Cross-Linguistic Differences

• Everyone has encountered a language that

contains a speech sound that is not in their native language

• As a general point, languages differ greatly

both in terms of how many phonemes they have and in terms of which phonemes they have

Being Distinctive

• We refer to the phonemes above as distinctive

because they make contrasts between

different words

• This can be illustrated for stops by using

minimal pairs: a pair of words that differ in

only one phoneme:

Lost in Translation

R vs L

• Many Korean and Japanese speakers have trouble learning

the contrast between R and L in English

• It is NOT because these sounds are absent in the native

language

– Seoul vs Korea

– Notice that they are NOT in contrastive positions (l at the

end of a syllable but r at the front)

– But English uses R and L contrastively: minimal pairs

– bLue~bRew, Light~Right, maLt~maRt

– Phonology is not just about the sound inventory, it’s also

about how sounds are put in use

Differences that are not

– But: the distinction between aspirated and

non-aspirated [p] is not distinctive in English (although it

is in other languages) That is, in English there are

no pairs like

[phIt] ‘hole in the ground, etc.’

[pIt] (whatever this might be)Rule of thumb: come up with minimal pairs

as a test for phonemes

Transcribing differences

• When we transcribe speech sounds using the

IPA notation, we may do so in different ways

– If we are interested in every phonetic detail,

we would indicate effects like aspiration in English, even if it is not distinctive ([phIt])

– If we are interested more in the

phonological inventory, we would omit the aspiration, as it is not distinctive ([pIt])

• For our purposes we will be concentrating on

the latter type

• Sometimes when we focus on phonology, an

abstract representation, we use slashes,

e.g /p/

Phonemes and Allophones

• Sometimes the same phoneme is pronounced

in different ways depending on its context

• The variants of a phoneme are called

allophones of that phoneme

• When we are talking about such distinctions,

the phoneme is in slashes /…/ and the

allophones are in square brackets […]

• The aspiration of e.g /p/ is a case of this type;

we say that /p/ in English has the allophones [p] and [ph]

Phonemes and Allophones,

cont.

• So, for instance, the phoneme /p/ appears in

each of the following words:

pit spit

• How, the first contains the allophone [ph],

while the second contains [p]

• In fact, the rule for aspiration in English is

more general & complex:

1 English voiceless stops (e.g., p, k, etc.) are

a. Aspirated if word initial, or syllable-initial

preceding a stressed vowel:

Compare récord vs recórd

b Otherwise unaspirated

Phonemes: Nasalized vowels

• e.g English speakers have not memorized any nasal vowels

• but English speakers do make nasal vowels: mat [mӕt]

Implications for learning words

• When we learn words, we don’t memorize their

pronunciations directly

• We memorize the abstract phoneme representations

(e.g., /pit/ and /spit/)

• The aspiration rule will turn /p/ in the former to an

aspirated [ph]

• This saves a tremendous amount of memory

but involves online computation

Phonemes and Allophones

allophone allophone

Finding Phonemes: More in

recitation

• The phonemes differ from language to language.

• How do we figure out what the phonemes of a language

are?

• One trick is to look for minimal pairs (pӕt bӕt)

Finding Phonemes

• Minimal pairs are two words that have different

meanings, but differ in only one sound sip/zip,

day/bay, ram/ran/rang

• Since the difference between the sounds is meaningful, it

must be stored in memory

• Our minimal pairs above let us conclude that:

– s/z are distinct phonemes,

– d/b are distinct phonemes,

– m/n/ŋ are distinct phonemes

Finding Phonemes

• Sometimes it isn’t possible to find minimal pairs for all

sounds, but speakers can tell whether a contrast would yield a distinct possible word, even if it’s not a real

word

• e.g “bat” vs “bap”: I know that “bat” is a word and

that “bap” isn’t, and that “bap” is a possible word

So /t/ /p/ are distinct phonemes

Rules of Pronunciation

• [vowel] ➔ [nasal] / [nasal consonant]

• the kind of sound that changes

• in this rule, it’s vowels

• Note: man but not mat

• Note: man and can and tan and san(k)

Rules of Pronunciation

• [vowel] ➔ [nasal] / [nasal consonant]

• “becomes”

Rules of Pronunciation

• [vowel] ➔ [nasal] / [nasal consonant]

• the change

• here, nasal

• Notice: I didn’t put [nasal vowel] I don’t have to put

vowel b/c that doesn’t change

• Only put what changes = simpler, less to remember

Rules of Pronunciation

• [vowel] ➔ [nasal] / [nasal consonant]

• “when”

Rules of Pronunciation

• [vowel] ➔ [nasal] / [nasal consonant]

• This is the environment that causes the change

• The underlining shows the position of the sound that’s

changing

• Here: before a nasal consonant

Rules of Pronunciation

• To show “after a nasal consonant”, we could have done

this:

• [nasal consonant] _

• To show “between a nasal consonant and a nasal

consonant”, we would have done this:

• [nasal consonant] _ [nasal consonant]

Rules of Pronunciation

• So, English speakers unconsciously know the following rule:

• [vowel] ➔ [nasal] / [nasal, consonant]

• “vowels become nasal when before a nasal consonant”

• This is a rule of assimilation, making sounds more similar.

• Rules of dissimilation (making sounds less similar) also

exist, but are less common

Phonemes and their

distributions

• Other cases of distinctive features lead to

some interesting observations

• Consider the nasals:

– rum run rung

• These phonemes are distinct at the end of the

word; but, [ng], unlike the other two, has the property that it never occurs word-initially in English

– map nap *ngap

• In order to understand these patterns, we

have to move from the phonemes to the

principles by which phonemes are organized into words and other units

The Syllable

• A familiar notion is that of the syllable: as in,

‘Philadelphia’ has five syllables

– Newborns perceive speech in terms of

syllables

– Basically, each vowel corresponds to a

syllable

• A refined set of hypothesis about the syllable

is important for many linguistic

generalizations

• Definitions (initial):

– Onset: the beginning of the syllable

– Nucleus: vowel in the middle of the syllable– Coda: consonant(s) at the end of a syllable

Syllable Structure

Legislator: le-gi-sla-tor; four syllables ( σ )

σ

Onset Rhyme k

Nucleus Coda

æ t

Monosyllabic cat:

Onsets and Speech Errors

Spoonerisms (Rev Dr W A Spooner, 1844-1930)

Target: d ear old qu een

Output: queer old dean

Target: You have w asted the whole t erm

Output: You have tasted the whole worm.

Target: You m issed my h istory lectures.

Output: You hissed my mystery lectures.

Further aspects of the

syllable

• Onset:

– English normally allows two consonants

– [s] can be added initially in many cases as

well, resulting in onsets with three

consonants (e.g splash)

– All sounds can occur in this position with

the exception of [ng] Thus the subdivision

of the syllable is crucial for stating this

generalization

• Coda:

– English normally allows two consonants,

although again there are cases where more

stack up (e.g belts)

Syllables and

well-formedness

• Conditions on syllable structure define a set of

(phonologically) possible words in a language;

for instance

– Actual words: brick, true, free, crab, etc.

– Non-words that are possible words of

English: blick, clee, flork

– Impossible words: *bnick, *fnee, *dmay

– Words in which historical change has made

an initial consonant silent: knee, knight,

gnat

– Another reason that we don’t just memorize

words but form generalizations over them

Differences across languages

• Languages differ in terms of the constraints

they impose on syllable structure:

– E.g Hawaiian:

• No coda consonants

• Maximum of one consonant per onset

• Examples: ink > 'înika

Norman > Nolemana

– E.g Polish: many consonant clusters at the

beginning of words that are impossible in English:

Infixation: more on this next week

Suffix: Attached to the end of a word (work- ed )

Prefix: Beginning ( un -important)

Infix: Inside a word

What is an example of an infix in English? There

is at least one phenomenon with the relevant

properties

this illustrates the basic principle that larger linguistic units are built out of smaller ones

Expletive infixation

Expletive Infixation is not something that our English teachers instruct us in; yet we know

a great deal about it – what’s the rule?

Go home and try with your friends & Tas

It has to do with stress patterns of language