Difference between revisions of "Phonotactics of Scottish Gaelic"
AndrewCarnie (talk | contribs) |
MicayaClymer (talk | contribs) (→Onsets) |
||
(12 intermediate revisions by the same user not shown) | |||
Line 5: | Line 5: | ||
===Onsets=== | ===Onsets=== | ||
− | We see onsets of the type ∅, C, CC, and CCC. Note that CCC onsets may only occur word-initially. Word-internal CCC clusters occur at syllable boundaries. Since the syllabification of polysyllabic Gaelic words is the subject of some debate (c.f. Clements 1985, Bosch 1998, etc.), I shall discuss word-initial and word-medial clusters seperately. | + | We see onsets of the type ∅, C, CC, and CCC. Note that CCC onsets may only occur word-initially. Word-internal CCC clusters occur at syllable boundaries. Since the syllabification of polysyllabic Gaelic words is the subject of some debate (c.f. Clements 1985, Bosch 1998, etc.), I shall discuss word-initial and word-medial clusters seperately. In word-initial C1C2 clusters, C2 is generally a member of the set /l, r, s/, and C1 is a sonorant or a stop (Wolters 1997). Various simplification processes work on these clusters. Clusters of the type #Cn are generally realized as #Cr, and . /hl/, /hn/, and /hr/ are frequently realized as /l/, /n/, and /r/, respectively (Gillies 1993). Word-internally, CC clusters are more varied because they often occur at syllable boundaries. Wolters (1997) notes that /s/ seems to have a special status in Gaelic, combining more freely in consonant clusters than other phonemes, even in cases where it violates the Sonority Sequencing Principle. |
− | |||
− | |||
− | In word-initial C1C2 clusters, | ||
− | |||
− | |||
− | Word-internally, CC clusters are more varied because they often occur at syllable boundaries. Wolters (1997) notes that /s/ seems to have a special status in Gaelic, combining more freely in consonant clusters than other phonemes, even in cases where it violates the Sonority Sequencing Principle. | ||
− | |||
Word-initial CCC clusters are of the form: | Word-initial CCC clusters are of the form: | ||
Line 18: | Line 11: | ||
{| class="wikitable" border="1" | {| class="wikitable" border="1" | ||
|- | |- | ||
− | |/s/ + C[+obs, -son, -voiced] + C [+son, -obs, +voiced] | + | |/s/ + C[+obs, -son, -voiced] + C [+son, -obs, +voiced] |
|} | |} | ||
Line 25: | Line 18: | ||
{| class="wikitable" border="1" | {| class="wikitable" border="1" | ||
|- | |- | ||
− | | ''sgrìobhadairean'' | + | | ''sgrìobhadairean'' |
− | | [skɾivət̪ɛɾɪn] | + | | [skɾivət̪ɛɾɪn] |
− | | " writers" | + | | " writers" |
|} | |} | ||
− | Word-internal CCC clusters result from coda-onset combinations and are therefore more varied, but they are often simplified by either deletion or epenthesis. For example, word-internal C1C2C3 clusters that begin with a nasal may be reduced to | + | Word-internal CCC clusters result from coda-onset combinations and are therefore more varied, but they are often simplified by either deletion or epenthesis. For example, word-internal C1C2C3 clusters that begin with a nasal may be reduced to ∅C2C3 with nasalization on the preceding vowel. (Wolters 1997). |
===Codas=== | ===Codas=== | ||
− | Codas may be of the type ∅, C, or CC. | + | Codas may be of the type ∅, C, or CC. Word-final CC clusters generally comprise a sonorant followed by an aspirated (voiceless) stop. |
− | |||
− | |||
− | Word-final CC clusters generally comprise a sonorant followed by an aspirated (voiceless) stop | ||
− | |||
− | |||
− | |||
− | |||
− | |||
− | |||
+ | In dialects that contain /ŋ/, it occurs only word-finally (Wolters 1997). Also, word-final nasals may be realized as syllabic segments (e.g. maduinn [mɑdn̩]; Ó Murchú 1988). | ||
Ó Murchú (1988) records several word-final CCC codas in stressed monosyllables, all containing /s/ (e.g. /xgs/), and notes that unstressed syllables seem to have a smaller inventory of possible codas than stressed syllables in the Perthshire dialect. | Ó Murchú (1988) records several word-final CCC codas in stressed monosyllables, all containing /s/ (e.g. /xgs/), and notes that unstressed syllables seem to have a smaller inventory of possible codas than stressed syllables in the Perthshire dialect. | ||
Line 50: | Line 35: | ||
==Phonological processes== | ==Phonological processes== | ||
===Assimilation effects=== | ===Assimilation effects=== | ||
− | Consonant clusters are affected by several types of assimilation effects, for example palatalization. In C1C2 sequences where C1 is a non-labial sonorant and C2 is any non-labial consonant, the consonants agree in terms of palitalization, even with an intervening epenthetic vowel. | + | Consonant clusters are affected by several types of assimilation effects, for example palatalization. In C1C2 sequences where C1 is a non-labial sonorant and C2 is any non-labial consonant, the consonants agree in terms of palitalization, even with an intervening epenthetic vowel. Labials may ocurr before either palatal and non-palatal consonants (Clements 1986). The overall effect is that most word-internal consonant clusters are either completely palatalized or not palatalized with nonhomogeneous clusters ocurring at word boundaries. Word-initial consonant clusters are generally [-palatal] (Wolters 1997). |
− | |||
− | |||
− | Labials may ocurr before either palatal and non-palatal consonants (Clements 1986). | ||
− | |||
− | |||
− | The overall effect is that most word-internal consonant clusters are either completely palatalized or not palatalized with nonhomogeneous clusters ocurring at word boundaries. | ||
− | |||
− | |||
− | Word-initial consonant clusters are generally [-palatal] (Wolters 1997). | ||
E.g.: | E.g.: | ||
Line 70: | Line 46: | ||
|} | |} | ||
− | In consonant clusters that contain a velarized sonorant, velarization assimilation often affects the other consonant(s). | + | In consonant clusters that contain a velarized sonorant, velarization assimilation often affects the other consonant(s). Finally, voicing assimilation also affects some consonant clusters. For example, a sonorant may be partially de-voiced before a voiceless stop, as in sult [suɫ̥t] ‘fat’, or an historically tense sonorant may result in partial voicing of a following stop, e.g. calltain [kɑɫdiɲ] ‘hazel’ (Wolters 1997). |
− | |||
− | |||
− | |||
===Aspiration=== | ===Aspiration=== | ||
Line 88: | Line 61: | ||
| "at" | | "at" | ||
|} | |} | ||
+ | |||
+ | Ladefoged (1998) noted differential realization of pre-aspiration, with /t/ and /k/ phonemes receiving greater pre-aspiration than /p/ and non-palatals more than palatals. | ||
===Vowel epenthesis=== | ===Vowel epenthesis=== | ||
Line 101: | Line 76: | ||
Word-final devoicing may often by overridden by phrasal effects and is most often seen phrase-finally (Wolters 1997). | Word-final devoicing may often by overridden by phrasal effects and is most often seen phrase-finally (Wolters 1997). | ||
− | + | ==Vowel quality== | |
− | Wolters (1997) argues that, while in many languages vowel quality influences consonant quality, in Scottish Gaelic, consonant quality influences vowel quality, especially w.r.t. palatalized (slender) consonants, with an association between palatal consonants and nearby [+front] vowels or on- or off-glides surrounding palatal consonants. | + | Wolters (1997) argues that, while in many languages vowel quality influences consonant quality, in Scottish Gaelic, consonant quality influences vowel quality, especially w.r.t. palatalized (slender) consonants, with an association between palatal consonants and nearby [+front] vowels or on- or off-glides surrounding palatal consonants. Vowels may also become nasalized in proximity to nasal consonants. For example, the word làmh (‘mother’) contains the putative nasal fricative and, according to Oftedal (1956) is realized in Leurbost Gaelic as [ɫãv]. |
+ | ===Diphthongs and hiatus=== | ||
+ | When two pronounced vowels are contiguous, we may see either a diphthong (monosyllabic) or a hiatus (disyllabic) (Ò Murchi 1988) | ||
+ | |||
+ | E.g.: | ||
+ | {| class="wikitable" border="1" | ||
+ | |- | ||
+ | | ''fiach'' | ||
+ | | [fiəx] | ||
+ | | "debt" | ||
+ | |- | ||
+ | | fitheach | ||
+ | | [fi(h)əx] or [fiʔəx] | ||
+ | | "raven" | ||
+ | |} | ||
− | + | In some dialects, the diphthong/hiatus distinction is erased or realized via differences in intonation. For example, O Murchu (1988) proposes a three-way classification system for complex nucleii, distinguishable by length and tone (rising or falling intonation). | |
==See Also== | ==See Also== |
Latest revision as of 15:44, 14 January 2010
Under construction!
Contents
Distribution of consonants
Onsets
We see onsets of the type ∅, C, CC, and CCC. Note that CCC onsets may only occur word-initially. Word-internal CCC clusters occur at syllable boundaries. Since the syllabification of polysyllabic Gaelic words is the subject of some debate (c.f. Clements 1985, Bosch 1998, etc.), I shall discuss word-initial and word-medial clusters seperately. In word-initial C1C2 clusters, C2 is generally a member of the set /l, r, s/, and C1 is a sonorant or a stop (Wolters 1997). Various simplification processes work on these clusters. Clusters of the type #Cn are generally realized as #Cr, and . /hl/, /hn/, and /hr/ are frequently realized as /l/, /n/, and /r/, respectively (Gillies 1993). Word-internally, CC clusters are more varied because they often occur at syllable boundaries. Wolters (1997) notes that /s/ seems to have a special status in Gaelic, combining more freely in consonant clusters than other phonemes, even in cases where it violates the Sonority Sequencing Principle.
Word-initial CCC clusters are of the form:
/s/ + C[+obs, -son, -voiced] + C [+son, -obs, +voiced] |
E.g.:
sgrìobhadairean | [skɾivət̪ɛɾɪn] | " writers" |
Word-internal CCC clusters result from coda-onset combinations and are therefore more varied, but they are often simplified by either deletion or epenthesis. For example, word-internal C1C2C3 clusters that begin with a nasal may be reduced to ∅C2C3 with nasalization on the preceding vowel. (Wolters 1997).
Codas
Codas may be of the type ∅, C, or CC. Word-final CC clusters generally comprise a sonorant followed by an aspirated (voiceless) stop.
In dialects that contain /ŋ/, it occurs only word-finally (Wolters 1997). Also, word-final nasals may be realized as syllabic segments (e.g. maduinn [mɑdn̩]; Ó Murchú 1988).
Ó Murchú (1988) records several word-final CCC codas in stressed monosyllables, all containing /s/ (e.g. /xgs/), and notes that unstressed syllables seem to have a smaller inventory of possible codas than stressed syllables in the Perthshire dialect.
Phonological processes
Assimilation effects
Consonant clusters are affected by several types of assimilation effects, for example palatalization. In C1C2 sequences where C1 is a non-labial sonorant and C2 is any non-labial consonant, the consonants agree in terms of palitalization, even with an intervening epenthetic vowel. Labials may ocurr before either palatal and non-palatal consonants (Clements 1986). The overall effect is that most word-internal consonant clusters are either completely palatalized or not palatalized with nonhomogeneous clusters ocurring at word boundaries. Word-initial consonant clusters are generally [-palatal] (Wolters 1997).
E.g.:
dragh | [traɣ] | "trouble" |
In consonant clusters that contain a velarized sonorant, velarization assimilation often affects the other consonant(s). Finally, voicing assimilation also affects some consonant clusters. For example, a sonorant may be partially de-voiced before a voiceless stop, as in sult [suɫ̥t] ‘fat’, or an historically tense sonorant may result in partial voicing of a following stop, e.g. calltain [kɑɫdiɲ] ‘hazel’ (Wolters 1997).
Aspiration
In #C1C2 clusters where C2 is a stop, it is generally unaspirated. Since the voicing contrast in Gaelic is generally realized as aspiration, the distinction between voiceless (aspirated) and voiced (unaspirated) consonants disappears post-consonantally (Wolters 1997).
Voiceless (aspirated) stops are also often realized as pre-aspirated post-vocalically (Ladefoged 1998).
E.g.:
aca | [aʰkə] | "at" |
Ladefoged (1998) noted differential realization of pre-aspiration, with /t/ and /k/ phonemes receiving greater pre-aspiration than /p/ and non-palatals more than palatals.
Vowel epenthesis
Most sonorant-obstruent clusters are broked up by an epenthetic vowel, realized as [ə] or colored by the surrounding vowels (Wolters 1997).
Homorganicity constraints
Sonorant-obstruent clusters are homorganic (at the same place of articulation). For example, we see [mb, nd, ŋɡ]. Non-homorganic sonorant-obstruent sequences are interrupted by an epenthetic vowel (Clements 1986).
Nasalization constraints
Nasals often have a nasalization effect on the syllable to which they belong. For example, vowels following the putative nasal fricative are often nasalized (Wolters 1997).
Word-final devoicing
Word-final devoicing may often by overridden by phrasal effects and is most often seen phrase-finally (Wolters 1997).
Vowel quality
Wolters (1997) argues that, while in many languages vowel quality influences consonant quality, in Scottish Gaelic, consonant quality influences vowel quality, especially w.r.t. palatalized (slender) consonants, with an association between palatal consonants and nearby [+front] vowels or on- or off-glides surrounding palatal consonants. Vowels may also become nasalized in proximity to nasal consonants. For example, the word làmh (‘mother’) contains the putative nasal fricative and, according to Oftedal (1956) is realized in Leurbost Gaelic as [ɫãv].
Diphthongs and hiatus
When two pronounced vowels are contiguous, we may see either a diphthong (monosyllabic) or a hiatus (disyllabic) (Ò Murchi 1988)
E.g.:
fiach | [fiəx] | "debt" |
fitheach | [fi(h)əx] or [fiʔəx] | "raven" |
In some dialects, the diphthong/hiatus distinction is erased or realized via differences in intonation. For example, O Murchu (1988) proposes a three-way classification system for complex nucleii, distinguishable by length and tone (rising or falling intonation).