Phonology II

Phonology II

Simple Rules and Metrical Phonology

Plan

• In the last class, we saw how the phonology of a language operates in terms of features that collectively make up speech sounds

• It is essential to remember that phonological operations make reference to features; in the first part of today’s class, we’ll look at how to treat such operations in terms of rules.

• In the second part of the class, we’ll look at another level of phonological organization, which concerns how sounds are organized into syllables and larger units.

Review: Alternations• In our last lecture, we saw alternations: cases in

which an object that is “one thing” at some level of analysis surfaces with distinct variants

• That’s a more precise way of saying that e.g. the English plural morpheme has some underlying form (I.e., /z/), but surfaces in different forms: /z/, /s/, /@z/

• This alternation is phonological in nature• Other alternations are not; consider that the

plural of box is boxes, but the plural of ox is oxen. This alternation is morphological nature.

Introducing phonological rules

• We need a way of formalizing these changes• As a basic assumption, we’ll assume that every word

and morpheme has an Underlying Representation, or UR.

• The UR is the basic sound form of that object; that is, how it is stored in the memory of speakers

• These UR’s are subject to operations-- rules-- that derive the surface forms of the words

Rules, abstractly

• In the basic form that we’ll be using here, rules have some sort of input that specifies what the rule applies to, and an output that shows the change that is effected

• Abstractly, these can be written as follows:X Y/ __ Z

• Informally, a rule of this type says that X (the input) is converted into Y (the output) when X is to the left of Z.

An example• Remember from above we saw some examples from Old

English in which there’s a single phoneme underlying [f] and [v] on the surface

• Let’s assume that this is underlyingly /f/• Recall the distribution: [v] occurs between vowels.• We can express this semi-formally as follows:

f v/ V__ V• This rule says that f becomes v between vowels• If we needed to be more precise about this, we could specify

this in terms of features. Then, what would change is a specification of the [voc] feature (I.e., the rule would change it from [-voc] to [+voc]

• This more precise rule would have to be written to voice /f/ intervocalically! Notice that if we are not careful, we might have a rule that voices too many consonants

Another example

• Let’s take an assimilation process• Consider (a subset of…) words prefixed with in- in

English; these express contradiction/negation:– inappropriate– intolerant– insincere

• In these forms, the idea is that we are adding a morpheme to express the negative part

• Based on the examples we see above, we would think that the nasal consonant in this morpheme is /n/

In-prefixation, continued

• Notice now that not all words with this prefix have the same pronunciation of that component:– impossible– imbalanced– impure

• In fact, the change in pronunciation is even reflected in the spelling (not all changes are!)

• What’s the pattern? We get the m-realization in front of labial consonants (labial stops)

Accounting for the pattern

• There are some more details about the in- prefix that we could look at; for the moment, think just about the facts we have seen

• We want to say that a nasal consonant becomes labial when a labial stop follows. One way of doing this:

C[+nas] C[+nas,+lab]/__ C[+obst,+lab]

Think about it…

The rule on the last slide shows how rules work. There’s actually more to be said about the facts of assimilation:

• Not all morphemes behave the same; compare unpopular with impossible…

• What about glides? I.e. unwelcome… are there any in- words like this?

• What about other consonants? How do people say e.g. incapable? What about unclear?

Types of rules: Summary• Phonological rules effect many different types of changes:

– Some make adjacent sounds more similar; these are often called Assimilation rules (remember e.g. English plurals)

– There are also Dissimilation rules, that make nearby sounds more distinct

– Some rules add a consonant or a vowel or a consonant that isn’t there in an Underlying Representation. These are called rules of Epenthesis.

• The most important thing in most cases is to find the environment in which the rule is triggered, so that you are stating the generalization about the environment correctly

Rule Interactions

• When linguists analyze the phonology of a language, they often find that many different rules apply to derive the surface forms

• That is, any given word that you see might be the result of more than one rule

• One difficult issue is that in many cases, rules have to apply to a form in a specific order

• This is kind of an advanced topic, but worth considering

Illustration

• Here’s a simplified example; some nouns in Polish, in singular and plural: sg pl trans ruk rog’i ‘horn’ lut lodi ‘ice’

• There are two things going on in these nouns:– One rule devoices voiced consonants at the end of words– Another rule raises /o/ to /u/ when a voiced consonant follows

• In order to get the correct results, the vowel raising rule has to apply before the consonant devoicing rule!

• Take home lessons: (1) there are some cases in which more than one rule has to be posited; (2) sometimes rule ordering is crucial

Syllables and Metrical Structure

• Transition: To this point, we have been looking at the smaller units in phonological analysis: features, and their arrangement into segments (along with the question of what the phoneme is)

• One large part of the phonology of a language concerns the orgranization of segments into larger objects: syllables.

• Syllable structure is relevant both for generalizations about segment distributions and for generalizations about stress (and related phenomena) in a language

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

‘Philadelphia’ has five syllables• 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

• As we will see in the next slide, these three do not count the same

Syllable Structure

A single syllable’s () structure:

Onset Rhyme k

Nucleus Coda æ t

Monosyllabic cat:

Comments

• Each of the different components of the syllable may be relevant for different generalizations (what goes into an onset versus what can appear in the coda)

• The nucleus and the coda form a unit together that excludes the onset

• The motivation for this arrangement has to do with the fact that onsets appear to be irrelevant for some aspects of syllable structure (e.g. nuclei and codas count together for syllable weight)

• Just focus on the different parts for the moment, not the motivation for this internal arrangement

Onsets and Speech Errors

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

Target: dear old queenOutput: queer old dean

Target: You have wasted the whole termOutput: You have tasted the whole worm.

Target: You missed my history lectures.Output: You hissed my mystery lectures.

Pig Latin

A language game showing sensitivity to structure:

Hat > AtheyThin > IntheyStrong > OngstreyLinguistics > InguisticsleyApple > Appleey

Pattern: Add a syllable to the end of the word; copy the onset of the word-initial syllable to the onset of the word-final; add ‘ey’ to the rhyme.

Further aspects of the syllable• Onsets:

– 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.

• The Nucleus:– Sometimes a consonant can function as a nucleus, like in kitten.

• The coda:– English normally allows two consonants, although again there are

cases where more stack up (e.g. belts)– Sometimes the morphology produces something that is hard to

pronounce: consider cases like the pronunciation of 3/8

Syllables and well-formedness• Conditions on syllable structure define a set of

(phonologically) possible words in a language; for instance:– Actual words, obstruent + liquid: brick, true, free,

crab, etc.– Non-words that are possible words of English:

blick, clee, flork– Impossible words, e.g. obstruent + nasal: *bnick,

*fnee, *dmay– Compare: words in which historical change has

made an initial consonant silent: knee, knight, gnat

Differences• 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:bzdura "nonsense"

babsk "witch"grzbiet [gzhbyet] "back" marnotrawstw [mar-no-trafstf] "of wastes"

Metrical Structure

Illustrating our tacit knowledge of abstract properties of language.

English Stress:indePENdenttwenty-SEvenLEgislatorMissisSIppi

Metrical structure, cont.• An active topic in phonological research concerns the stress

patterns of words. • It appears that the world’s languages group syllables into

feet, like the iamb and the trochee:– Iamb: (’) – Trochee (’)

• You will have seen something similar before in the study of metrical verse (poetry); one member of the foot is prominent (stressed).

• Languages differ in terms of what types of feet they use, whether they involve iterative feet throughout a word, and so on.

• English stress turns out to be sort of complicated for stress for a variety of reasons, but we can see some aspects of foot structure nevertheless.

The Rhythm Rule

Remember those words we saw above?

TWEnty-seven MISsissippi LEgislators

• I.e., without having been taught this rule, speakers automatically calculate a shift in stress

•Not a rule like that coming from English teachers (e.g. “No split-infinitives”); part of what speakers know unconsciously

•Consider e.g. thirteen Japanese bamboo tables

Expletive Infixation

We’ll do morphology next week; for fun now:

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.

Patterns

Expletive Infixation is not something that ourEnglish teachers instruct us in; yet we know a great deal about it:

1) *fanta-f*cking-stic fan-f*cking-tastic

2) unrea-f*cking-listic *unreali-f*cking-stic

*unre-f*cking-alistic

• *indepen-f*cking-dent inde-f*cking-pendent

Abstract Metrical StructureStress placement involves computation over abstractrepresentations, and these define expletive infixation.

Recall: grouping into feet; one syllable in the foot receives prominence (stress):

Àppalàchicóla

(Àppa)(làchi)(cóla)

E.g.

Has three binary feet ():

In which the first member of each foot is prominent.

The RuleThe placement of the expletive is governed bythe following condition:Expletive Insertion: The expletive is inserted immediately before a non-initial stress foot of the word:

(Appa)(lachi)(cola) 1 2

Has two such junctures; so eitherAppa-f*cking-lachicola, or Appalachi-f*cking-cola

(Disclaimer: the rule is actually more complicated…)