Science of Language Week 3

What is a syllable?

A syllable is a grouping of sound(s) composed of a central peak of sonority (usually a vowel) the consonants that lie around this central peak The permissible combinations of segments in syllables are language-specific (phonotactics). What are the sounds you're allowed/not allowed to put together ("phonotactically illegal"/"legal")

Which sound doesn't belong?

For the sets of phonemes below, determine (a) which one does NOT belong and (b) what is the natural class that remains once it is eliminated? [v ð z f ʒ] all fricatives, mostly voiced, except for: a) [f] b) voiced fricatives 2. [m b g p w] all labial except: a) [g] b) labials 3. [m ŋ j n] all nasals except: a) [j] b) nasals 4. [l r t] all liquids except: a) t b) liquids 5. [u o e ʊ ɔ] all back rounded vowels except: a) [e] b) back rounded vowels 6. [f s ʔ θ h ʃ] all continuant except: a) [ʔ] b) unvoiced continuants

Phonetics

Signal (acoustic signal/articulation)- production of sounds Details about sounds

Free variation

Some allophones are in free variation: They can appear in the same environment or position in the word. Speaker can use one or another allophone, but it is at their discretion: free variation FREE TO VARY when you use one or the other E.g., the word pop can sound different by "exploding" the final /p/ or closing your lips tightly, leaving the /p/ unreleased. You may use one variant or the other depending on your personal speaking style, formal versus informal context, etc.

Features define natural classes

Sounds can be grouped into natural classes based on shared distinctive features Can reflect shared manner, voicing, place of articulation, or some combination Major class distinction: Vocalic vs. consonantal sounds [+vocalic] sounds produced with the vocal tract mostly open [+consonantal] sounds produced with the articulators blocking airflow (partly or completely)

Some examples of natural classes

Stops, fricatives, affricates A.k.a. obstruents [+consonantal, -sonorant] Fricatives [+consonantal, -sonorant, +continuant] Voiceless stops [+consonantal, -sonorant, -continuant, -delayed release, -voice] When you describe a natural class, it must be ALL and ONLY the sounds in the language that fit that set of features. Fricatives:use features that describe fricatives and only fricatives

Nucleus

center, meat, vowel may be diphthong

Allophones are in

complementary distribution

Different environment ≡

complementary distribution ⇒ allophones

Phonemes are in

contrastive distribution

Same environment ≡

contrastive distribution ⇒ (distinct) phonemes

Coda

tautosyllabic consonants after the nucleus consonants after vowel but within the same syllable end/closing portion of syllable (may be in the middle of a word)

Onset

tautosyllabic consonants before the nucleus before the vowel but within the same syllable tautosyllabic = within the same syllable

Phonology

"sign" (abstract properties of sound structure) - stored in your head/what you already know- how sounds pattern within a language Abstract groupings of sounds into classes Rule-governed patterning of sounds in a language

4 Key concepts for "doing" phonology:

1. Phoneme vs. allophone 2. Natural classes 3. Features (used to define natural classes) 4. "conditioning environment" (what is the environment that causes a sound to change into a different allophone?)

Doing phonology problems

6 STEPS: Write down the environments in which the sounds occur. (Make two lists.) Examine those environments to determine whether they are the same or different. Determine whether sounds are contrastive (occur in the same environment) or complementary (occur in different environments). Phoneme = same environment Allophone = different environment If allophones, state a generalization about the distribution of sounds (see next slide) where do we find one sound and the other?

Allophone vs phoneme- phoneme analogy

Are Superman and Batman distinct (phonemes) or are they "allophones"? Step 1: Determine the environment. Where do they each occur? Both occur in dangerous situations Step 2: Examine the environment. Are the environments the same or different? Same environment- both in dangerous environments

Other important features

Are affricates /ʧ, ʤ/ [+continuant] or [-continuant]? They are [-continuant] because there is cessation of oral airflow during production of sounds They are distinguished from stops by the feature [+delayed release]. Certain fricatives ([s, z, ʃ, ʒ]) and the affricates ([ʧ, ʤ]) are characterized by particularly loud high- frequency noise, classified as [+strident, +coronal] features (sibilants as a class, strident as a feature) needs to be distinguished as +coronal because there are other strident sounds that are not coronal (f,v) The /l/ sound is distinguished by the feature [+lateral]: produced with airflow along the sides of the mouth. Nasals are distinguished by the feature [+nasal] because air flows in the nasal cavity.

Allophone vs phoneme- analogy

Are two sounds different phonemes or allophones of the same phoneme? If two things are phonemes, they are different, and that difference is important allophones mean two versions of the same thing, and they appear in different environments take Clark Kent/Superman as an example CLARK KENT/SUPERMAN = ALLOPHONES Where does each occur? Superman appears only in dangerous situations. Clark Kent appears only in nondangerous situations (everywhere else). That means they are in complementary distribution (i.e., do not overlap, never occur in the same environment). They are really the "same person," i.e., they are allophones of the same phoneme.

Anatomy of a syllable

Greek letter sigma (σ) used for the syllable C and V: used as shorthand for "consonant" and "vowel" when describing allowable syllable shapes in a language E.g., CV, CCV, CCCV

If the two sounds are allophones

If the two sounds are allophones, we want to be able to describe where they occur. Example: English aspiration (simplified)

Features intro

If two sounds contrast phonemically, they must differ by one or more distinctive features. We have talked about pairs of voiced and voiceless sounds, like /p/ and /b/. Produced with the vocal folds vibrating: [+voice] Produced without vocal fold vibration: [-voice] /p/ and /b/ contrast by a single distinctive feature, [voice]. Other features for consonants describe phoneme place of articulation and manner of articulation. voicing is the simplest one Binary: one or the other (2 options)

Important features: vowels

Important Features: Vowels: (7 vowel features) [-consonantal] [+syllabic] (differentiates vowels from glides)- can be the core of a syllable [high] [low] [back] [round] [tense]

Phonological knowledge: legal clusters

Languages also differ in which sounds can occur in consonant clusters, and in what order. Russian words can start with clusters like /mn/ and /stv/ that are not possible in English. Phonotactics ≡ Restrictions on possible phoneme sequence allowable/not allowable sequence of phonemes in a syllable Restrictions typically differ for onset versus coda sequences: Onset-only clusters: br-, tw-, kl-, str-, sl-, sn-... Coda-only clusters: -ld, -lf, -nd, -mp...

Phonological knowledge:syllable shapes

Languages differ in what syllable shapes are permitted. Hawaiian allows one onset consonant and no coda consonants (e.g., kalikimaka, "Christmas"). CV syllables only English allows up to three onset consonants and four coda consonants (e.g., sprints, twelfths). Up to CCCV(C), (C)VCCCC 3 consonants in onset, 4 in coda Polish allows up to four onset and five coda consonants (e.g., pstry "gaudy," nastepstw "consequences" (gen.pl.) Up to CCCCV(C), (C)VCCCCC

Syllables and phonological rules

Many phonological rules/processes make reference to where a sound occurs within a syllable. Aspiration in english: when something is word initial, will be aspirated- what's relevant is that it's syllable initial, and needs to be the onset of a stressed syllable Some dialects of English (British, Boston, New York) delete /r/ in certain contexts. Where? Have to talk about syllable position- /r/ is deleted in coda, not onset position This rule is sensitive to syllable structure /r/ is deleted when it occurs in the coda but not when it occurs in the onset.

Minimal pair

Minimal pair: Two words that differ by only one sound

Allophones

One version of a sound- same phoneme but sound is different in different contexts- not enough to have a change in meaning Now compare with too vs. stew. [thu] vs. [stu] [th] and [t] are different versions of the same sound. One is aspirated and one is not. "allophones of the same phoneme" Sometimes one allophone occurs in one particular phonetic environment and the other allophone always occurs in a different context We say that these allophones are in complementary distribution Some t's are aspirated and occur in certain places, and some t's are not aspirated and will occur in different places, but the two t's are complementary to each other Some allophones are in free variation: They can appear in the same environment or position in the word.

Onset Maximization Principle

Onset should have as much stuff in it until it's no longer allowed, and the rest should go into the coda If there is more than one way to syllabify segments across a syllable boundary, choose the one that maximizes the number of segments in the onset without forming illegal clusters (way to check: can a word start with this cluster?) E.g., "instruct" = in.struct not i.nstruct (nstr is illegal as an onset) can have a word that starts with nstr not ins.truct (not maximal onset) tr is legal onset, but haven't put everything possible in not inst.ruct (not maximal onset) r is allowable, but haven't put everything possible in not instr.uct (not maximal onset) 3 consonants maximum in onset

Open vs closed syllable

Open: does not have a coda ("boy") Closed: has a coda consonant

Phonemes

Phonemes are the minimal unit of sound used to represent lexical contrast. If I have one sounds and I substitute a different sound and I get a different word, that is a lexical contrast or a phoneme tip vs. dip Changing from one to the other changes the meaning of the word. Therefore, /t/ and /d/ are phonemes (they are the minimal units of contrast) Minimal pair: Two words that differ by only one sound We say that two phonemes are in contrastive distribution: Distribution means where do the sounds occur- can create a contrast in meaning The distribution of phonemes create a contrast

Features

Phonemes ≡ "bundles" of distinctive features stop p = set of features that make that sound (unvoiced, bilabial) Features are the building blocks of phonology Can be combined in different ways to create the phonemes of a language Typically binary: (two versions- either there or not there) Either present ([+feature]) or Absent ([-feature]) in a given phoneme E.g., [± voice], [± round]

Phonological rules

Phonological rules are typically defined with respect to features, syllable/word boundaries Rule: X → Y/ A___B Translation: x becomes y in the context of ___ E.g., voiceless stop consonants are aspirated word initially (and syllable initially) Aspiration: [+spread glottis] [+cons, -cont, -voice -delayed release](voiceless stops) → [+spread glottis] (become aspirated) /#___ # = word initial position

Syllabification

Transcribe the word. Identify the syllable nuclei (and build up the syllable tree). Add onsets using the onset maximization principle. Then add leftover consonants to codas.

What phoneme is it?

What Phoneme is It?: Phonemes can be depicted as matrices of distinctive features, as seen below. OK to omit redundant features (e.g., all vowels are voiced, so we don't include [+voice] for vowels). What phoneme is it?: +consonantal -sonorant +continuant +voice +labial /m/ /g/ /v/ /b/

Syllabification

Why was the /m/ in "limit" in the onset of the second syllable and not in the coda of the first syllable? Why [lɪ.mɪt], and not [lɪm.ɪt]? Native speakers have reasonable intuitions about the number of syllables in a word but terrible intuitions about the location of syllable boundaries. What principles govern where the syllable boundary lies?

natural classes

[+sonorant]: Vowels, glides, liquids, nasals Have a resonant property Sonorants are vowels and vowel-like consonants. They are inherently voiced in english (no voiced-voiceless contrast) and involve smooth rather than turbulent (noisy) airflow [-sonorant] sounds are the obstruents: stops, fricatives, affricates [+continuant]: Air can escape through the (center) of the oral cavity Fricatives, liquids, glides [-continuant]: stop the flow of air in the oral cavity Stops, affricates, nasals (constriction in oral cavity even though air is flowing out of nasal cavity)

Hierarchical organization

[klin] ("clean") Not simply "beads on a string" - each phoneme next to the one before it Different structure of how the sounds that are next to each other are grouped and whether that's a closer grouping than sounds on the other side of it ∃ hierarchical organization of the segments How the segments get grouped into smaller units within the larger unit of the syllable

natural classes for place

[labial]: Produced with one or both lips Bilabials, labiodentals [coronal]: Produced with the tip or blade of tongue Interdentals, alveolars, postalveolars [dorsal]: Produced with the back of the tongue Velars

Data set

[phəl] "fruit" [kəl] "yesterday" [pəl] "moment" [kapi] "copy" [khəl] "wicked person" [kaphi] "ample" Are [p, k] and [ph, kh] distinct phonemes, or are they allophones of the same phoneme? List environments. Examine distribution.

Rhyme

nucleus + coda vowel plus closing consonant think words that rhyme: "tan, pan, stan, man"- nucleus and coda are the same in all of those

English data set: aspiration

pan [phæn] span [spæn] tone [thon] stone [ston] key [khi] ski [ski] First, list the environments Looks like the environments are different. Therefore, we say the two sounds are in "complementary distribution." Now we want to be able to say WHERE one occurs and WHERE the other occurs. How would you describe the two environments? The aspirated environment is in word initial position before a vowel, and the unaspirated environment is after the consonant [s] and before a vowel.

What syllables are allowed in the language =

phonotactics