Transcript Diachronic Consonantal Phonotactics in Latin

The Work of Syllable Structure or Linear Sequence?
Ranjan Sen
University of Oxford
2
Phonotactics: Two Approaches
Syllable Approach
 Range of contrasts in an
environment attributed to
position within syllable
Linear Approach
 Range of contrasts in an
environment attributed to
linear segmental sequence
alone
Which approach tackles best the diachronic phonotactic
development seen in the history of Latin?
3
Phonotactic Relevance of the
Syllable in Latin
 Notions “well-formed onset” and “well-formed coda”
required in syllabification
•C: any
•CC: stop (or /f/) + liquid
•s- extrasyllabic
•C: any
•CC: sonorant + voiceless stop
•-s extrasyllabic
 Word-based Syllable Hypothesis : iːn.síg.nis ‘notable’
supported by accent-placement, but /g/ not found
word-finally, whereas /gn/ found word-initially
(gnaːrus ‘having knowledge of’)
4
Voice Assimilation
 Regular regressive assimilation in biconsonantal
sequences (C1C2)
 Stop + stop: *scriːb-to-s > scriːptus ‘written’, obtinuiː ‘I
obtained’ = [pt], e.g. OPTINVI
 Stop + fricative: *nuːbsiː > nuːpsiː ‘I married’
 Fricative + stop: *is-dem > *izdem > iːdem ‘same’
 Every obstruent in a consonantal sequence agrees in voice
regardless of syllabification
 Regardless of syllabification: plebs ‘people’ = [pleps]
5
Place and
Frication
= regressive place and frication assimilation
C2
C1
The Place Hierarchy:
Dorsal > Labial > Coronal
Stop C1 lower than or level
with C2 on hierarchy
assimilates to C2 in place and
frication
Syllable Approach: “codas
stops unspecified for coronal
place regardless of the
environment, and labial place
if followed by dorsal stop”
 clearly unsatisfactory: no
motivation for recourse to
syllabic position – linear
sequence is necessary and
sufficient
Dor
Dor
Lab
Cor
*ec-ce >
ecce ‘look!’
ec-pːonoː ‘I
bring out’ (=
expoːonoː)
lact-is ‘milk
(gen.)’
dorsal
+ fricative (no
fricative)
Lab
*ob-kaidoː >
occiːdoː ‘I
knock
down’
dorsal
+ fricative (no
fricative)
Cor
*hod-ce >
*hocce >
hoc ‘this
(neut.)’
dorsal
+ fricative (no
fricative)
ec-feroː
(Plautus) ‘I
carry out’
*deik-siː >
[diːksiː] ‘I
said’
*ob-petoː >
oppetoː ‘I
optimus
‘best’
encounter
prematurely’
opi-ficiːna >
*opficiːna >
officiːna
‘workshop’
*quid-pe >
quippe ‘for’
ad-feroː >
afferoː ‘I
deliver’
*nuːb-siː >
nuːpsiː ‘I
married’
*pat-tos >
*patsos >
passus
‘suffered’
*quat-siː >
quassiː ‘I
shook’
6
Manner
The Manner Hierarchy:
(for place assimilation)
Fricative > Stop > Nasal
Fricative C1
• before fricative C2
only
• obeying Place
Hierarchy
• *disfacilis >
difficilis ‘difficult’
Stop C1
• before C2 of any
manner
• obeying Place
Hierarchy
• *quidpe > quippe
‘for’; adferoː >
afferoː ‘I deliver’,
*kaidmentom >
cae(m)mentum
‘rubble’
•
Nasal C1
before any
obstruent C2
regardless of Place
Hierarchy
• before nasal C1
obeying Place
Hierarchy
• *kemtom >
centum ‘hundred’,
*in-maneoː >
immineoː ‘I
overhang;
threaten’ vs.
autumnus
‘autumn’
7
Nasality
= no nasal or place assimilation
= regressive nasal assimilation
= regressive nasal and place assimilation
• Nasal C2 nasalises stop C1, which
also assimilates in place to C2
obeying Place Hierarchy
• Exception: failure of nasal
assimilation in Dor + /m/
 Again, Syllable Approach
unsatisfactory
• Better starting-point: linear
configuration Dor + /m/
• Cf. early epenthesis: Greek
dráchma  drac(h)uma ‘Greek
coin’, tegmen > tegimen/tegumen
‘covering’
Nasal C2 Lab
Cor
C1
Dor
*sekmentom > *deknos >
segmentum = dignus =
[gm] ‘piece’
[ŋn] ‘worthy’
Lab
*supmos >
summus
‘highest’
Cor
*kaidmentom > *atnos >
*caimmentum annus ‘year’
> caementum
‘rubble’
*swepnos >
somnus
‘sleep’
8
Hypothesis – Linear Sequence
Feature x, if poorly cued relative to adjacent more
robustly cued feature, is neutralised and assimilated to
adjacent more robustly cued feature
 External cue: release into vowel, thus C2 features usually
more robustly cued than C1 features
 Internal cue: Place Hierarchy – Dor > Lab > Cor
 Internal cue: Manner Hierarchy for place feature –
Fricative > Stop > Nasal
9
Scale for occurrence of contrasts
More contrasts
Before
vowel
Before
liquid
• Voice
• Place
• Voice
• Place
• Manner
• Manner
(with
exception)
• all with
exceptions
Fewer contrasts
Fricative
before
nasal/
stop
• No
voice
• Place
• Manner
Other Stop
Obstr
C
before before
before stop fricative
nasal
• No
• No
voice
voice
• No
voice
• Place if
• Place if
place
place
hierarchy
hierarchy
admits
admits
• Place if
place
hierarchy
admits
• Manner if
place
hierarchy
admits
• No
manner
ex. Dor
+ /m/
Nasal
before
obstr
• No
voice
• No
place
• Manner
10
•Sonorants appear before C2 of any voice specification
comparoː : combiboː
verpa : verbum
 sonorants unspecified for voice pre-consonantally
•Nasal C2 triggers voicing of C1: *sekmentom > segmentum
 nasals voice-specified pre-vocalically
•Liquid C2 allows voice contrast in C1: capra : criːbrum
 liquids unspecified for voice pre-vocalically
•BUT /s/ > [z] post-vocalically before voiced consonant,
including liquids: *preslom > [prezlom] > preːlum
 liquids voice-specified pre-vocalically??
11
Sonorant Voice Specification
 Voicing of /r/ at early stage
 Early merger in Latin of /sr/ and [ðr] inherited from Proto-Italic:
*fuːnesris > fuːnebris ‘funereal’
 Voiced epenthetic stop before /r/ vs. voiceless epenthetic stop
before /l, n/:
*gheimrinos > *heimbrinos > hiːbernus ‘wintry’ vs. *exemlom >
exemplum ‘example’, autumnus > autumpnus ‘autumn’
 /l,m,n/ became voice-specified later (in archaic period)
 /s/ before /l,m,n/ > [z] (with consequent loss of [z] + compensatory
lengthening of the preceding vowel):
*preslom > preːlum ‘wine-/oil-press’, cosmis > coːmis ‘friendly’,
*casnos > caːnus ‘white(-haired)’
 voice-specified? capra vs. criːbrum?
12
Return of the Syllable:
Why does liquid C2
allow preceding
voice contrast if
voice-specified?
TR Onsets
 Unspecified if
in stop + liquid
onset (not σ-initial)
Phonetically
based:
incline vs.
ink-like
 Divergent syllabifications of identical sequence:
 *po.plos > populus ‘people’
 *pop.li.kos > poblikos > puːblicus ‘public’
13
Morphological Pressures
More thoroughgoing
regressive
assimilation in
prefix + verbal root
contacts
More faithful
retention of root
shape elsewhere in
verbal morphology
Some morpheme
boundaries
conditioned syllable
boundaries, thus
determining the
voicing of sonorants
*sub-regoː >
surrigoː ‘I
rise’ vs.
eːbrius
‘drunk’
*sum-to-s >
*sumptus
‘assumed’
vs. *kemtom
> centum
‘hundred’
*nek-legoː >
neg.le.goː ‘I
neglect’ vs.
Aisclaːpius,
poːclum
14
Latin phonotactic development driven by linear
segmental sequence, not syllable structure
Relevant parameters include internal factors
(manner and place hierarchies) and external
factors (release features, coarticulatory cues)
Syllables relevant in distribution of voice in
sonorants
Morphology could directly override phonetic
considerations in phonotactics, and indirectly via
syllable structure
Only direct influence of syllable structure in Latin
phonotactics is in determining what onsets/codas
are well-formed
Diachronic Phonotactic Development in Latin
Ranjan Sen
Linguistics, Philology and Phonetics
University of Oxford, U.K.
[email protected]