phonological strength: background

CONSONANTS 4

Phonological strength

Phonological strength: Background

phonological theory: what is possible in phonology? what is a

possible phonological process? + why?

2 types of phonological process:

- determined by (linear) adjacency / locality (interaction with a

segment’s (sub)segmental neighbourhood): primarily,

assimilation (+ dissimilation, etc.): green + peas =

gree[mp]eas, Hu. zseb ’pocket’ + kendő ’tissue’ = zse[pk]endő

’hankie’

- ???

??? = lenition/weakening

Lenition theory investigates the types of consonantal processes that

cannot be easily reduced to assimilation, etc.

what is possible in the patterning of a lenition process

Phonological strength

weak = frequent site for lenition (weakening, see below)

strong = more resistant to lenition (stability, or less weakening

than in weak position – implicational hierarchy)

certain positions (e.g., stressed vowels, word-initial or post-coda

consonants) are stronger than others (e.g., unstressed vowels,

word-final or coda consonants) both synchronically and

diachronically

universal tendencies + parameters (e.g., pre-consonantal vs.

final codas, word-initial C, stress-sensitivity, quality of C1 for

post-coda C2, etc., see Scheer & Ségéral 2008)

Several strands of phonological theory have focused on phenomena

known as lenition (see, e.g., Lass & Anderson 1975, Lass 1984,

Harris 1994, 1997, Kirchner 1998, Ségéral & Scheer 1999, 2008,

etc.)

there are two ways in which lenition has been claimed to be

phonologically interesting:

i. the set of phonological processes involved

ii. the set of environments in which lenition can/can’t occur

some work tries to link the two, to provide a fully explanatory model

(cf., e.g., Lass & Anderson 1975, Lass 1984, Harris 1994, 1997,

Kirchner 1998, Ségéral & Scheer 1999, 2008, etc.)

We focus here on (ii), and simply assume that some link in terms of

(i) applies, grouping together all and only those processes which

‘weaken’ segments

(i) the processes:

classical definition (Vennemann, recorded in Hyman 1975: 165): “a

segment X is said to be weaker than a segment Y if Y goes through

an X stage on its way to zero”, this links different segment types in

terms of their relative “weakness” on a continuum (sometimes

called a “hierarchy of phonological strength” or “lenition

trajectory”), along which segments are assumed to be likely to

progress diachronically, e.g.:

A lenition trajectory (Lass 1984: 178):

... which allows us to group together processes of different types as

cases of lenition, e.g.:

spirantisation (a segment becoming a fricative)

sonorisation (i.e., the loss of obstruency, e.g., tapping/flapping in

accents of English)

delaryngealisation (i.e., the loss of [voice] in so-called voice

languages, traditionally called devoicing, as in a number of Slavic

languages)

‘voicing’ (i.e., lenisisation in so-called aspiration languages like

English)

spirantisation (a segment becoming a fricative)

e.g., medial d-spirantisation, Middle English period, cca. 1400:

(+ further examples from Liverpool, see later)

sonorisation (i.e., the loss of obstruency)

e.g., tapping/flapping in accents of English:



languages)

Polish:


English)

aspiration (or: spread glottis) languages base the traditional

voiceless--voiced distinction on the presence vs. absence of

aspiration (rather than on vocal chord vibration)

cf.:

Hungarian English

/t/ no vocal chord vibration

no aspiration

no vocal chord vibration

aspiration

/d/ vocal chord vibration

no aspiration


no aspiration


English)




cf.:

Hungarian English

/t/


no aspiration

voiceless


aspiration

voiceless aspirated

/d/

vocal chord vibration

no aspiration

voiced


no aspiration

voiceless unaspirated


English)




cf.:

English

/t/


aspiration

voiceless aspirated

FORTIS

/d/


no aspiration

voiceless unaspirated

LENIS

cf.



languages)

and


English): fortis -> lenis, i.e., the loss of fortisness/aspiration

cf.



languages)

and


English): fortis -> lenis, i.e., the loss of fortisness/aspiration

both are cases of lenition/weakening!

lenisisation accompanying Scouse (Liverpool) diddification: (hypocoristics, diddy ‘small’ cf. diddly ‘nothing’)

(cf. Australian English: Australian → Aussie / Ozzie)

lenisisation accompanying Scouse (Liverpool) diddification: (hypocoristics, diddy ‘small’ cf. diddly ‘nothing’)

(cf. Australian English: Australian → Aussie / Ozzie) http://www.youtube.com/watch?v=VlfQDQ6KWFs

Example 1: lenition in Liverpool English

lenition trajectory:

cf.

Example 2:

(ii) the environments:

Environmental patterning is another thing that lenitions have in

common (indeed, some argue that it is all that they have in

common).

What is usually assumed in lenition taxonomies (e.g., Scheer and

Ségéral 2008):

there are a possible set of strong positions and a possible set of

weak positions

individual cases of lenitions in languages must obey this

implicational hierarchy: if they occur in a strong position, they

must also occur in a weak position

that is:

The ‘stress parameter’:

lenition only occurs if the vowel following the segment is not

stressed, means that there can be two types of medial

environment (cf. esp. Harris 1994, 1997 for English):

“intervocalic” needs to be split into:

a stronger [ v__V ] and a weaker [ V__v ] (the latter:

“ambisyllabic”)

How does English fit into this system?

Find examples of strong, weak1, weak2.

lenition in English1: typical symptoms:

strong: stressed/full vowel, aspiration, /h/

weak V: reduced (typically: schwa), syncope

weak1 C: (pre)glottalization, unreleased plosive, no /h/

weak2 C: tapping/flapping, no /h/2

1 By "English", I mean the major tendencies constituting a common core of the phonologies of the

varieties of English. 2 For the Two directions for lenition, see Szigetvári (1999, 2008).

lenition sites: a typical representative of the Germanic pattern:

word-initial C is strong

stress-sensitive system in its consonants: pre-stress vs. post-stress

behaviour

consonants are strong before full vowels (atomic, vehicular) but

weak after them (cf. atom and vehicle)

zero-stressed syllables contain schwas (or syllabic consonants –

irrelevant here)

only schwas are subject to deletion (syncope – battery vs. batt'ry)

Ideally, models of lenition need to not overgenerate and be able to

account for only those environments in which lenition is possible.

But they must not undergenerate, either.

recall:

phonological theory: what is possible in phonology? what is a

possible phonological process? + why?

therefore: refine the taxonomy of

strong(er) and weak(er) positions!

Refinement #1:

The “Withgott effect”

tapping/flapping: the ‘classical’ pattern: roughly, in

intervocalic position whenever the second vowel is

unstressed



unstressed

but: Withgott (1982): tap suppression in certain positions

(for certain speakers):

flapped t aspirated t

capitalístic militarístic

sanitisátion

monotonícity

what is the difference between these

words?



unstressed

but: Withgott (1982): tap suppression in certain positions

(for certain speakers):

flapped t aspirated t

capitalístic militarístic

sanitisátion

monotonícity

cf. capital vs. military, sanitise, monotone: untapped t in the

derivative where there is untapped t in the base due to stress

on the syllable whose onset the t is

[Steriade (2000: 322-326):

paradigm uniformity (PU) conditions: promote invariance of

some sound property within a

given paradigm (= the set of

morphologically related words):

Paradigm Uniformity

All surface realizations of μ, where μ is the morpheme shared

by the members of paradigm x, must have identical values for

property P. (Steriade 2000: 313)

tap suppression in militaristic is a PU effect]

but: the Withgott effect is also found in morphologically

simple examples, e.g.:

Mediterránean, Winnipesáukee, Navratilóva,

abracadábra, etc.:

aspiration (instead of lenition)

but: no other member in the paradigm

therefore:

the problem of the third syllable in a dactyl:

capitalístic vs. militarístic

foot-based solution: cyclic analysis:

(cápital) (míli)(tary) + -(ístic)

(capita)(lístic) (mili)(ta(rístic))

+ adjunction of the stray syllable to the right: (Navra)(ti(lóva))

etc. (Withgott 1982, Jensen 2000, Davis 2003, 2005)

foot-initial plosives are aspirated



(Navra)(ti(lóva))

Expletive infixation data seem to support these footings

(Davis 2003): the expletive element is inserted between feet

for emphasis, cf.:

Winne-frickin-pesaukee and Winnepe-frickin-saukee

mili-fuckin'-taristic and milita-fuckin'-ristic

but capita-frickin-listic and *capi-frickin-talistic



(Navra)(ti(lóva))



(Navra)(ti(lóva))

also works for potato (y)

recall: foot-initial plosives are aspirated!

the foot-based solution:

accounts for the Withgott effect

accounts for expletive infixation data

no need to refer to PU effect -> accounts for underived

examples (e.g., Navratilova) as well

also accounts for initial unstressed syllables (e.g., potáto)

the foot-based solution:

accounts for the Withgott effect

accounts for expletive infixation data

no need to refer to PU effect -> accounts for underived

examples (e.g., Navratilova) as well

also accounts for initial unstressed syllables (e.g., potáto)

BUT: only applicable to nonfinal dactyls!

next:

final dactyls are also special (cf. vánity, compétitive)

right adjunction is ruled out: nowhere to adjoin

the only way is to the left… :-P -- ??

phonological strength: background

Documents