Neutralizations to h in Mösiehuali from the perspective of Cognitive Grammar David Tuggy SIL-Mexico.

Neutralizations Neutralizations to to hh in in

MösiehualMösiehualiifrom the perspective from the perspective

of Cognitive Grammarof Cognitive Grammar

David TuggyDavid TuggySIL-MexicoSIL-Mexico

Neutralizations to Neutralizations to hh in in MösiehualiMösiehuali

MösiehualMösiehualii, or Tetelcingo Nahuatl, is , or Tetelcingo Nahuatl, is noteworthy for the number of noteworthy for the number of contrasting phonemes which are contrasting phonemes which are neutralized in certain syllable-final neutralized in certain syllable-final positions.positions.

In particular, a syllable-final In particular, a syllable-final hh may, may, by alternation evidence, be shown to by alternation evidence, be shown to be a be a pp, , tt, , kk, , kʷkʷ, , c̸�c̸�, , čč, , ss, , šš, , ww, or , or yy

not to mention an not to mention an hh..

Neutralizations to Neutralizations to hh in in MösiehualMösiehualii

An An hc̸hc̸ ̸̸ sequence, for instance, could sequence, for instance, could correspond to any of the following:correspond to any of the following: hc̸�hc̸� ((ayoh-c̸�īn-ƛiayoh-c̸�īn-ƛi ‘zucchini/young squash’, ‘zucchini/young squash’, ayoh-ƛi ayoh-ƛi

‘squash’)‘squash’) tc̸tc̸ �� ( (otimomah-c̸�īnootimomah-c̸�īno ‘you H knew yourself’ vs. ‘you H knew yourself’ vs. otimomatotimomat

‘you knew yourself’)‘you knew yourself’) c̸�c̸�c̸�c̸� ( (otimokeh-c̸�īnootimokeh-c̸�īno ‘you H stood’ vs. ‘you H stood’ vs. otimokec̸�otimokec̸� ‘you ‘you

stood’)stood’) čc̸�čc̸� ( (nēh-c̸�ahc̸�ilianēh-c̸�ahc̸�ilia ‘he calls me’ vs. ‘he calls me’ vs. nēč-ilwia nēč-ilwia ‘he tells me’)‘he tells me’) sc̸�sc̸� ( (mo-nemilih-c̸�īmo-nemilih-c̸�ī ‘your H life’ vs. ‘your H life’ vs. mo-nemilismo-nemilis ‘your life’) ‘your life’) wc̸�wc̸� ( (mo-kʷah-c̸�īmo-kʷah-c̸�ī ‘your H stick’ vs. ‘your H stick’ vs. kʷaw-iƛ kʷaw-iƛ ‘stick’)‘stick’) yc̸�yc̸� ( (mo-kʷēh-c̸�īmo-kʷēh-c̸�ī ‘your H skirt’ vs. ‘your H skirt’ vs. kʷēy-iƛkʷēy-iƛ ‘skirt’) ‘skirt’)

Neutralizations to Neutralizations to hh in in MösiehualMösiehualii

Similarly, a final Similarly, a final øø (zero) can correspond (zero) can correspond to an to an hh, a , a ww, a , a yy, a nasal, or a zero., a nasal, or a zero.hh ( (mo-ayomo-ayo ‘your squash’ vs. ‘your squash’ vs. ayoh-ƛi ayoh-ƛi

‘squash’)‘squash’)w w ((mo-kʷa mo-kʷa ‘your stick’ vs. ‘your stick’ vs. kʷaw-iƛkʷaw-iƛ ‘stick’) ‘stick’)yy ( (mo-kʷēmo-kʷē ‘your skirt’ vs. ‘your skirt’ vs. kʷēy-iƛkʷēy-iƛ ‘skirt’) ‘skirt’)nn ( (mo-c̸�omo-c̸�o ‘your hair’ vs. ‘your hair’ vs. c̸�on-ƛic̸�on-ƛi ‘hair’) ‘hair’)mm ( (mo-kōmo-kō ‘your waterpot’ vs. ‘your waterpot’ vs. kōm-iƛkōm-iƛ

‘waterpot’)‘waterpot’)øø ( (mo-konēmo-konē ‘your baby’ vs. ‘your baby’ vs. mo-konē-c̸�īmo-konē-c̸�ī

‘your H baby’)‘your H baby’)

Usage-based structureUsage-based structure

Cognitive Grammar brings several Cognitive Grammar brings several interesting theoretical perspectives to interesting theoretical perspectives to data of this kind.data of this kind.

In the first place, it is “usage based”In the first place, it is “usage based” Words and other forms which are learned by Words and other forms which are learned by

speakers are speakers are ipso fac̸toipso fac̸to part of the language, part of the language, regardless of whether or not they might be regardless of whether or not they might be

predicted by rules or patterns (predicted by rules or patterns (sc̸hemassc̸hemas) ) which are also part of the language.which are also part of the language.

This has consequences for the way CG This has consequences for the way CG handles categorization.handles categorization.

CategorizationCategorization A quick review, or as the case may be, a A quick review, or as the case may be, a

crash course in how CG handles crash course in how CG handles categorization:categorization: Categorization is based on judgements of Categorization is based on judgements of

similarity from standard to target of similarity from standard to target of comparison.comparison.

If the standard is fully realized in the target If the standard is fully realized in the target this is full schematicity: S this is full schematicity: S T. T.

T “is an”/ “is a kind of” S.T “is an”/ “is a kind of” S. T is a sub-case of ST is a sub-case of S T “elaborates” ST “elaborates” S S is schematic for TS is schematic for T S is a generalization of which T is a specific caseS is a generalization of which T is a specific case

CategorizationCategorization When the standard is only partially realized in When the standard is only partially realized in

the target you have partial schematicity: S - - the target you have partial schematicity: S - - T.T.

T is similar to S, but is not an S.T is similar to S, but is not an S. T is a kind of warped or distorted S.T is a kind of warped or distorted S. T is an extension of S.T is an extension of S.

Typically, S is moreTypically, S is moreprominent (salient) than T prominent (salient) than T (we compare less-known T (we compare less-known T to well-known S.)to well-known S.)

When S - - When S - - T, what T, what is common to them is common to them (typically) forms an (typically) forms an incipient higher-level incipient higher-level schema.schema.

S T

S'

Usage-based structureUsage-based structure

The subcases prompt the extraction of The subcases prompt the extraction of the schema.the schema.

They do not disappear when it is They do not disappear when it is extracted, but continue to support it.extracted, but continue to support it.

General patterns are based on and General patterns are based on and learned from specific cases.learned from specific cases.

This is not to deny, however, that it works This is not to deny, however, that it works the other way also: a well-established the other way also: a well-established pattern pattern sanc̸tionssanc̸tions (legitimizes) specific (legitimizes) specific cases that match it.cases that match it.

If these are novel cases, the schema is If these are novel cases, the schema is used productively.used productively.

General patterns coexist General patterns coexist with with

specific structuresspecific structures Bottom line: both general patterns and Bottom line: both general patterns and

specific cases of those patterns can specific cases of those patterns can coexist in speakers’ minds.coexist in speakers’ minds.

They do not contradict but rather They do not contradict but rather reinforce each other.reinforce each other.

It is an empirical issue, not a It is an empirical issue, not a theoretically predetermined one, to theoretically predetermined one, to what extent any given pattern or what extent any given pattern or specific case is entrenched or prominent specific case is entrenched or prominent in speakers’ minds.in speakers’ minds.

CategoriesCategories

The more elaborations (sub-cases) a The more elaborations (sub-cases) a schema has, the better entrenched, schema has, the better entrenched, and thus more and thus more salientsalient, it is likely to be. , it is likely to be.

E2E1

SE3

E4

E5


Often sub-cases can support different Often sub-cases can support different schemas; i.e. they embody more schemas; i.e. they embody more than one pattern or generalization.than one pattern or generalization.

E2E1

SE3

E4

E5

S2S3


““Classical” categories, of the kind assumed Classical” categories, of the kind assumed by many theories, are modeled whenby many theories, are modeled when

(a) a salient (well-entrenched) pattern or (a) a salient (well-entrenched) pattern or schemaschema

(b) has many less salient sub-cases, and(b) has many less salient sub-cases, and (c) any other schemas based on those sub-(c) any other schemas based on those sub-

cases can be reasonably ignored. As a cases can be reasonably ignored. As a result,result,

(d) The category does not overlap with (d) The category does not overlap with neighboring categories.neighboring categories.


Well-behaved “classical” categories look like this:Well-behaved “classical” categories look like this: Or, even better, like this:Or, even better, like this: But “classical categories” are just one case of But “classical categories” are just one case of

the range of kinds of categories we find in the range of kinds of categories we find in language.language.

Schema1

Subcase1b Subcase1c

Subcase1a Subcase1n

Schema2

Subcase2b Subcase2c

Subcase2a Subcase2n

Schema3

Subcase3b Subcase3c

Subcase3a Subcase3n

Schema3Schema2Schema1

Undistinguished mass of cases, none standing

out from the others


out from the others


out from the others

nooverlap

nooverlap

Phonological CategoriesPhonological Categories

Phonological categories are of the same kinds Phonological categories are of the same kinds as any other linguistic categories.as any other linguistic categories.

Phonology, like semantics, lexicon or grammar, Phonology, like semantics, lexicon or grammar, is built bottom-up, and is based on usage.is built bottom-up, and is based on usage.

As with semantics, that usage is conditioned by As with semantics, that usage is conditioned by the fact that it occurs in bipolar, symbolic the fact that it occurs in bipolar, symbolic structures (i.e. in morphemes or grammatical structures (i.e. in morphemes or grammatical constructions.)constructions.)

Subcaseb Subcasec

Subcasea Subcasen

Schema

Schema


out from the othersE2E1

SE3

E4

E5

S2S3

nēč-nēč- in Mösiehual in Mösiehualii

MösiehualMösiehualii speakers learn or hear many word/stems speakers learn or hear many word/stems in which in which nēčnēč before a verb stem means “me”. before a verb stem means “me”.

Some are very well-entrenched, others less so.Some are very well-entrenched, others less so. These cases support a prefixal These cases support a prefixal nēč-nēč-V schema.V schema.

nēčihta

(3p) SEE me

nēčteriksa

(3p) KICK me

nēčneki

(3p) LOVE me

nēčasi

(3p) REACH me

nēčƛahčīwia

(3p) BEWITCH me

nēč …

(3p) DO X to me .

nēčmalīna

(3p) TWIST me

nēč-nēč- in Mösiehual in Mösiehualii

It is only within this It is only within this nēčnēč-V construction -V construction that that nēč-nēč- is identified as a morpheme. is identified as a morpheme.

Its prefixality consists in its strong Its prefixality consists in its strong attachment to the construction, in attachment to the construction, in which it is followed by a verb stem.which it is followed by a verb stem.

It can hardly be activated without It can hardly be activated without activating the construction. activating the construction.

nēč …

(3p) DO X to me .

Verb stem schema

nēh-nēh-

MösiehualMösiehualii speakers also learn or hear many word/stems speakers also learn or hear many word/stems in which in which nēhnēh before a verb stem also means “me”. before a verb stem also means “me”.

Again, some are very well-entrenched, others less so.Again, some are very well-entrenched, others less so. These cases support a pair of prefixal These cases support a pair of prefixal nēh-nēh-V schemas.V schemas.

nē ak̫ ilia

(3p) HOLD me BACK

nē atilia

(3p) DEAFEN me

nē ikalwia

(3p) STICK it onto me

nēhčōktia

(3p) MAKE me CRY

nēhčia

(3p) AWAIT me

nēhčihkopilia

(3p) WINK at me

. nēh …

(3p) DO X to me .

. nēh č…

(3p) DO X to me .

nēh-nēh-

(In Mösiehual(In Mösiehualii the maximal syllable is CVC. the maximal syllable is CVC. Thus in all these cases the Thus in all these cases the hh is syllable- is syllable-final.)final.) (In fact the vast majority of (In fact the vast majority of hh’s in the language ’s in the language

are syllable-final.)are syllable-final.)

nē ak̫ ilia

(3p) HOLD me BACK

nē atilia

(3p) DEAFEN me

nē ikalwia

(3p) STICK it onto me

nēhčōktia

(3p) MAKE me CRY

nēhčia

(3p) AWAIT me

nēhčihkopilia

(3p) WINK at me

. nēh …

(3p) DO X to me .

. nēh č…

(3p) DO X to me .

nēh-nēh-

These two schemas cry out to be These two schemas cry out to be compared to each othercompared to each other

The result of doing so is a higher-level The result of doing so is a higher-level schema.schema.

. nēh …

(3p) DO X to me .

. nēh č…

(3p) DO X to me .

(3p) DO X to me .

. nēh t[sib]…

nēč-nēč- alternates with alternates with nēh-nēh-

All three of these schemas cry out to All three of these schemas cry out to be compared with the better-be compared with the better-entrenched entrenched nēčnēč-V schema.-V schema.

. nēh …

(3p) DO X to me .

. nēh č…

(3p) DO X to me .

(3p) DO X to me .

. nēh t[sib]…nēč …

(3p) DO X to me .


(For convenience, we will concentrate on the (For convenience, we will concentrate on the comparison of comparison of nēč-nēč-V with the highest-level V with the highest-level nēhnēh-V schema.)-V schema.)

This is the Cognitive Grammar equivalent of This is the Cognitive Grammar equivalent of a morpheme alternation rule: a morpheme alternation rule: nēč-nēč- ~ ~ nēh-nēh- / / __ __ tt[sib][sib]

(Remember, it does not deny the existence (Remember, it does not deny the existence or importance of the more specific rules or importance of the more specific rules nēč-nēč- ~ ~ nēh-nēh- / __ / __ c̸�c̸� and and nēč-nēč- ~ ~ nēh-nēh- / __ / __ č č ) )

(3p) DO X to me .

. nēh t[sib]…nēč …

(3p) DO X to me .


This comparison of This comparison of nēč-nēč-V with the V with the nēh-nēh-V V schemas can again result in the extraction of a schemas can again result in the extraction of a schema.schema.

Like many other high-level schemas, this one Like many other high-level schemas, this one has gotten too abstract to be useful for most has gotten too abstract to be useful for most purposes, and is unlikely to be very salient.purposes, and is unlikely to be very salient.

(3p) DO X to me .

. nēh t[sib]…

(3p) DO X to me .

nēC …[-vd]

nēč …

(3p) DO X to me .


It is less than fully useful becauseIt is less than fully useful because(a)(a) It is not the case that just any voiceless It is not the case that just any voiceless

consonant can appear as the final part of the consonant can appear as the final part of the morpheme meaning ‘me’. morpheme meaning ‘me’.

(b)(b) To say an important part of the same thing: To say an important part of the same thing: čč and and hh do not form a particularly natural class. do not form a particularly natural class.

(3p) DO X to me .

. nēh t[sib]…

(3p) DO X to me .

nēC …[-vd]

nēč …

(3p) DO X to me .

tēč-tēč- alternates with alternates with tēh-tēh-

Through exactly parallel steps, speakers arrive Through exactly parallel steps, speakers arrive at a at a tēč-tēč-V schema and a V schema and a tēhtēh-V schema, and -V schema, and compare them to each other. compare them to each other.

So their cognitive grammars will have another So their cognitive grammars will have another morpheme alternation rule: morpheme alternation rule: tēč-tēč- ~ ~ tēh-tēh- / __ / __ tt[sib][sib]

(And also the more specific rules (And also the more specific rules tēč-tēč- ~ ~ tēh-tēh- / / __ __ c̸�c̸� and and tēč-tēč- ~ ~ tēh-tēh- / __ / __ č č ) )

(3p) DO X to us .

. tēh t[sib]…tēč …

(3p) DO X to us .

ƛapeč-ƛapeč- alternates with alternates with ƛapeh-ƛapeh-

Again, through largely parallel steps, Again, through largely parallel steps, speakers will come up with speakers will come up with ƛapečƛapeč and and ƛapehƛapeh- - tt[sib][sib] schemas (cf. schemas (cf. mo-ƛapečmo-ƛapeč ‘your bed’ and ‘your bed’ and mo-ƛapeh-c̸�ī mo-ƛapeh-c̸�ī ‘your H bed’,) and will compare ‘your H bed’,) and will compare them to each other. them to each other.

Another morpheme-alternation rule, Another morpheme-alternation rule, ƛapečƛapeč ~ ~ ƛapeh ƛapeh / __ / __ tt[sib][sib], will be established. , will be established.

ƛapeč (…)

BED

ƛapeh t[sib]

…BED (?)

čč alternates with alternates with hh

Such morpheme alternations will function as Such morpheme alternations will function as specific cases for the extraction of another specific cases for the extraction of another schema, which localizes the alternation to the schema, which localizes the alternation to the phonological segments which contrast. phonological segments which contrast.

tēč …

(3p) DO X to us .

(3p) DO X to us .

. tēh t[sib]…nēč …

(3p) DO X to me .

(3p) DO X to me .

. nēh t[sib]…

ƛapeč (…)

BED

ƛapeh t[sib]

…BED (?)

(…) h (t[sib]…)

(X)

(…) č (…)

(X)


This schema is the Cognitive Grammar This schema is the Cognitive Grammar equiva-lent of a phonological equiva-lent of a phonological (morphophonemic) rule (morphophonemic) rule č č h h / __/ __ t t[sib] [sib]

It can be abbreviated by leaving out the It can be abbreviated by leaving out the specifications, largely empty at this level of specifications, largely empty at this level of schematicity, of the semantic poles schematicity, of the semantic poles involved.involved.

(…) h (t[sib]…)

(X)

(…) č (…)

(X)

(…) h (t[sib]…) (…) č (…)


Nonetheless it is bipolar (morphemic) in that Nonetheless it is bipolar (morphemic) in that the comparison of phonological structures is the comparison of phonological structures is

motivated by and established largely because of motivated by and established largely because of the identity of the semantic structures. the identity of the semantic structures.

If If čč and and hh weren’t the only points of difference in weren’t the only points of difference in allomorphs, speakers probably wouldn’t particularly allomorphs, speakers probably wouldn’t particularly compare them, much less entrench the comparison compare them, much less entrench the comparison as part of their standard linguistic repertoire. as part of their standard linguistic repertoire.

Other similar abbreviated representations we use Other similar abbreviated representations we use should be understood to also be morphemic, not should be understood to also be morphemic, not just phonological.just phonological.

(…) h (t[sib]…) (…) č (…) (…) h (t[sib]…)

(X)j

(…) č (…)

(X)j

Other obstruents Other obstruents alternate with alternate with hh

Similar kinds of considerations, includingSimilar kinds of considerations, including many specific structuresmany specific structures comparisons among themcomparisons among them abstraction of schemas and comparisons abstraction of schemas and comparisons

among themamong them lead to the establishment of a number of lead to the establishment of a number of

other phonological rules similar to the other phonological rules similar to the čč -- hh rule. rule.

(…) h (t[sib]…) (…) č (…)

Other obstruents Other obstruents alternate with alternate with hh

They include: They include: a a c̸�c̸� -- hh schema (this is the most similar to the schema (this is the most similar to the čč --

hh rule) rule) a a tt -- hh schema (before schema (before tt or apical affricates) or apical affricates)

an an šš -- hh schema (before schema (before čč))

an an ss -- hh schema (before schema (before c̸c̸ ̸̸))

a a kk -- hh schema (before schema (before kk or or kʷkʷ))

(…) h (t[sib]…) (…) č (…)

(…) h (t[sib]…) (…) (…)

(…) h (t(x)…) (…) t (…)

(…) h (č…) (…) š (…)

(…) h ( …) (…) s (…)

(…) h (k(w)…) (…) k (…)

DissimilationDissimilation

These are all so similar that they invite These are all so similar that they invite comparison and extraction of a schema.comparison and extraction of a schema.

Such a schema would say an obstruent Such a schema would say an obstruent ((OO) is pronounced ) is pronounced hh when it precedes a when it precedes a similar obstruent.similar obstruent.

(…) h (t[sib]…) (…) č (…) (…) h (č…) (…) š (…)

(…) h (t[sib]…) (…) (…) (…) h ( …) (…) s (…)

(…) h (t(x)…) (…) t (…) (…) h (k(w)

…) (…) k (…)

(…) h (Oi-k …) (…) Oj (…)


This is in a sense a rather important This is in a sense a rather important schema.schema.

It captures the generalization that all It captures the generalization that all these patterns are dissimilations.these patterns are dissimilations.

(…) h (t[sib]…) (…) č (…) (…) h (č…) (…) š (…)

(…) h (t[sib]…) (…) (…) (…) h ( …) (…) s (…)

(…) h (t(x)…) (…) t (…) (…) h (k(w)

…) (…) k (…)

(…) h (Oi-k …) (…) Oj (…)


However, it cannot be the whole story. However, it cannot be the whole story. It does not tell you how similar the second obstruent It does not tell you how similar the second obstruent

need be in which cases.need be in which cases. It is the lower-level schemas which give you that It is the lower-level schemas which give you that

information.information. They themselves got it from the specific cases, of course.They themselves got it from the specific cases, of course.

(Tuggy 2004)(Tuggy 2004)

(…) h (t[sib]…) (…) č (…) (…) h (č…) (…) š (…)

(…) h (t[sib]…) (…) (…) (…) h ( …) (…) s (…)

(…) h (t(x)…) (…) t (…) (…) h (k(w)

…) (…) k (…)

(…) h (Oi-k …) (…) Oj (…)

pp alternates withalternates with kk and and hh

Note that there is no schema dissimilating Note that there is no schema dissimilating pp before another before another pp.. This is so even though both This is so even though both pp’s and ’s and hphp sequences sequences

occur in Mösiehualoccur in Mösiehualii, and the overall Dissimilation , and the overall Dissimilation schema in some sense predicts that there ought to schema in some sense predicts that there ought to be an extension linking them.be an extension linking them.

But in a usage-based model it is absent because But in a usage-based model it is absent because there simply are no clear cases where the there simply are no clear cases where the morphemic patterns place a morphemic patterns place a pp before another before another pp. .

There is thus some expectation (but no There is thus some expectation (but no requirement) that if such were to arise, the requirement) that if such were to arise, the combination would be pronounced combination would be pronounced hphp..

(…) h (Oi-k …) (…) Oj (…)

(…) h (p…) (…) p (…)


Syllable-final Syllable-final pp’s should occur when a verb like ’s should occur when a verb like kʷepakʷepa ‘turn’ has its final vowel truncated.‘turn’ has its final vowel truncated.

(This happens in preterite, durative and other forms) (This happens in preterite, durative and other forms) Instead of the expected Instead of the expected kʷepkʷep, the form , the form kʷekkʷek appears. appears.

(…) …VCV (…)

DO Process(pres, impf, fut, etc.)

(…) …VC (•#,C…)

DO Process(pret, dur, etc.)

kik̫ epa

(3p) TURN it OVER

(3p) TURNED it OVER

(3p) IS TURNING it OVER

okik̫ ekkik̫ ektika


Again, a schema can be extracted which captures this Again, a schema can be extracted which captures this pp

- - kk pattern. (Though abbreviated, it too is pattern. (Though abbreviated, it too is morphemic in nature.)morphemic in nature.)

This pattern would predict —but again, fail to guaranteeThis pattern would predict —but again, fail to guarantee— that a — that a pppp combination would be pronounced combination would be pronounced kpkp rather than rather than hphp..

(…) …VCV (…)

DO Process(pres, impf, fut, etc.)

(…) …VC (•#,C…)

DO Process(pret, dur, etc.)

kik̫ epa

(3p) TURN it OVER

(3p) TURNED it OVER


okik̫ ekkik̫ ektika

(…) k (• …) (… •) p (V…)

(…) k (p…) (…) p (…)


When When kʷepakʷepa is truncated and the following consonant is truncated and the following consonant is a is a kk, the truncated form is pronounced , the truncated form is pronounced kʷehkʷeh. .

The extensions from the more common truncated form The extensions from the more common truncated form kʷekkʷek follow the pattern of the follow the pattern of the kk - - hh dissimilation dissimilation rule. rule.

kik̫ epa

(3p) TURN it OVER

okik̫ ek

(3p) TURNED it OVER

kik̫ ektika


okik̫ ehki

(3p pl) TURNED it OVER

(…) h (k(w)…) (…) k (…)

Rule orderingRule ordering

This is an example of what “rule ordering” (in this This is an example of what “rule ordering” (in this case, feeding) looks like in Cognitive Grammar.case, feeding) looks like in Cognitive Grammar.

One extension leads to or allows for another.One extension leads to or allows for another. Each is a subcase of at least one more-general Each is a subcase of at least one more-general

pattern (schema, rule)pattern (schema, rule) This imposes a certain logical ordering on the rules.This imposes a certain logical ordering on the rules.

k̫ epa

k̫ ek

(…) …VC (•#,C…) (…) …VCV (…)

(…) k (• …) (… •) p (V…)

(…) h (k(w)…) (…) k (…)

Rule orderingRule ordering

Note that this ordering configuration does not Note that this ordering configuration does not preclude a direct extension from preclude a direct extension from kʷepakʷepa to to kʷehkʷeh, and , and thus from thus from p p to to hh..

It does show how the alternation is motivated by the It does show how the alternation is motivated by the way it fits with a couple of established patterns.way it fits with a couple of established patterns.

It is a much less surprising alternation in MösiehualIt is a much less surprising alternation in Mösiehualii than it would be in most other languages.than it would be in most other languages.

k̫ epa

k̫ ek

(…) …VC (•#,C…) (…) …VCV (…)

(…) k (• …) (… •) p (V…)

(…) h (k(w)…) (…) k (…)

SpirantizationSpirantization

In parallel ways to those we have seen In parallel ways to those we have seen in the case of dissimilation, a schema in the case of dissimilation, a schema for semivowel Spirantization is for semivowel Spirantization is extracted.extracted.

SpirantizationSpirantization

Specific forms like Specific forms like kʷawkʷaw and and kʷahkʷah, or , or kʷēykʷēy and and kʷēhkʷēh, are compared, and , are compared, and schemas extracted.schemas extracted.

k̫ awiƛ

STICK / WOOD /

TREE (thing)

k̫ ahme

GROUP

of stick/tree S

mok̫ ī

your STICK/TREE/

WOOD hon

k̫ ēyiƛSKIRT (thing)

mok̫ ē īyour SKIRT hon

(…) h ( •C…) (…) w (V…)

(…) h ( •C…) (…) y (V…)

Various consonants are Various consonants are pronounced pronounced hh

The two schemas The two schemas ww -- hh and and yy -- hh support the support the

semivowel (semivowel (WW) ) -- hh schema. schema. This schema can be compared with the Dissimilation This schema can be compared with the Dissimilation

schema and yet another, higher-level schema abstracted.schema and yet another, higher-level schema abstracted. This highest schema expresses the generalization that a This highest schema expresses the generalization that a

lot of consonants are pronounced lot of consonants are pronounced hh before other before other consonants.consonants.

(…) h ( •C…) (…) w (V…) (…) h ( •C…) (…) y (V…)

(…) h ( •C…) (…) W (V…) (…) h (Oi-k …) (…) Oj (…)

(…) h ( •C…) (…) C (V…)


Again, we have a schema that captures an Again, we have a schema that captures an important generalization, but is far from absolute.important generalization, but is far from absolute.

It does not tell us which consonants are It does not tell us which consonants are pronounced pronounced h h in which circumstances. Many in which circumstances. Many consonants are not pronounced as consonants are not pronounced as hh in many in many preconsonantal positions.preconsonantal positions. E.g. nasals, E.g. nasals, ll and and rr are not (cf. are not (cf. nonānc̸�īnonānc̸�ī ‘my mother’, ‘my mother’,

īmpaīmpa ‘on them’ ‘on them’ mokalc̸�īmokalc̸�ī ‘your hon. house’) ‘your hon. house’) E.g. E.g. kk is not before apicals, apicals beforeis not before apicals, apicals before kk or or pp, , šš

before before c̸�c̸�, etc., etc. And so forth.And so forth.

(…) h ( •C…) (…) W (V…) (…) h (Oi-k …) (…) Oj (…)

(…) h ( •C…) (…) C (V…)


The lower-level schemas give us that infor-The lower-level schemas give us that infor-mation; and they got it from the specific cases.mation; and they got it from the specific cases.

The whole system is based on specific usages. The whole system is based on specific usages. As they change, the system changes with As they change, the system changes with

them.them.

(…) h ( •C…) (…) W (V…) (…) h (Oi-k …) (…) Oj (…)

(…) h ( •C…) (…) C (V…)

PhonemesPhonemes Like most if not all languages, MösiehualLike most if not all languages, Mösiehualii has a has a

set of contrasting sounds that tend to be set of contrasting sounds that tend to be structured close to the norm for classical structured close to the norm for classical categories. categories.

They are abstracted from millions of individual They are abstracted from millions of individual usage events.usage events.

They tend to consist of a single articulatory They tend to consist of a single articulatory position or gesture, though sometimes a complex position or gesture, though sometimes a complex one.one.

They are of two main types: syllable peaks, and They are of two main types: syllable peaks, and syllable margins (onsets or endings).syllable margins (onsets or endings).

Among them they cover the range of articulatory Among them they cover the range of articulatory gestures of the language pretty completely, with gestures of the language pretty completely, with little overlap.little overlap.

They are, of course, phonemes.They are, of course, phonemes. We have assumed them in what we have been We have assumed them in what we have been

describing so far.describing so far.

A “classical” phonemeA “classical” phoneme

For example, the phoneme For example, the phoneme tt consists largely of a consists largely of a group of sounds which, though differing slightly group of sounds which, though differing slightly among each other, have in common a (non-nasal) among each other, have in common a (non-nasal) apico-dental gesture at the margin of a syllable apico-dental gesture at the margin of a syllable (and its acoustic correlates).(and its acoustic correlates).

Although specific pronunciations differ in detail, Although specific pronunciations differ in detail, none of those details particularly stand out. none of those details particularly stand out.

This is, then, a good example of a “classical” This is, then, a good example of a “classical” category.category.

(apico-dental stop)

t

(…) tə ( •C…)

(…V) t * (#) (… •) t (V…)

(…e) t (c̸…)=


out from the others

(apico-dental stop)

t

A “classical” phonemic A “classical” phonemic systemsystem

tt fits in with a number of other similar categories in a fits in with a number of other similar categories in a system which pretty well covers the territory of system which pretty well covers the territory of sounds pronounced at syllable margins. sounds pronounced at syllable margins. (Only the (Only the obstruents are represented here.)obstruents are represented here.)

(apico-dental stop)

t

Undistinguished subcases

(bilabial stop)

p


(velar stop)

k


(labialized

velar stop)

k̫


(alveolar

affricate)


(lateral alveolar

affricate)

ƛ


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(palato-alveolar

fricative)

š



These sounds, by and large, contrast clearly with each other, These sounds, by and large, contrast clearly with each other, each potentially occurring between vowels, for instance. each potentially occurring between vowels, for instance.

This looks like a good “classical” categorial system. If the This looks like a good “classical” categorial system. If the phonemes didn’t overlap, it would be perfect.phonemes didn’t overlap, it would be perfect.

(apico-dental stop)

t


(bilabial stop)

p


(velar stop)

k


(labialized

velar stop)

k̫


(alveolar

affricate)


(lateral alveolar

affricate)

ƛ


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(palato-alveolar

fricative)

š


nooverlap

nooverlap

nooverlap

nooverlap


Schemas can be extracted to group or categorize these Schemas can be extracted to group or categorize these phonemes, but they are not likely to be as useful or phonemes, but they are not likely to be as useful or prominent as the phonemes themselves are.prominent as the phonemes themselves are. These are CG counterparts to the traditional “archiphonemes”, or These are CG counterparts to the traditional “archiphonemes”, or

in some cases to phonological features or feature combinations.in some cases to phonological features or feature combinations.

(apico-dental stop)

t


(bilabial stop)

p


(velar stop)

k


(labialized

velar stop)

k̫


(alveolar

affricate)


(lateral alveolar

affricate)

ƛ


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(palato-alveolar

fricative)

š


velar O

grave Oapical O

A messy phonemeA messy phoneme

Not all phonemes present such a pretty picture, however.Not all phonemes present such a pretty picture, however. For example, the phoneme For example, the phoneme ww in Mösiehual in Mösiehualii includes a range of includes a range of

pronunciations including [pronunciations including [ww], [], [w+w+ ], [], [ββ], and [], and [ɸɸ].]. These “allophones” are clearly salient in people’s minds, to These “allophones” are clearly salient in people’s minds, to

such an extent that /such an extent that /ww/’s unity can be (and certainly has been) /’s unity can be (and certainly has been) questioned.questioned.

Yet when the allophones are compared a schema is extractible.Yet when the allophones are compared a schema is extractible. They clearly have a lot in common, there is alternation They clearly have a lot in common, there is alternation

evidence to link them together, and they do not contrast with evidence to link them together, and they do not contrast with each other (in native words, anyway).each other (in native words, anyway).

(… •) w (V…)[+bk]

(… •) β (V…)[-bk]

(… •) ɸ (V…)[-bk]

{ }h

(… •) (V…)[+bk]

{ }h

(labial and/or velar

approximant/fricative)


In fact, there are other sub-generalizations that In fact, there are other sub-generalizations that can (and should) be included in the picture.can (and should) be included in the picture.

(The relations between these sub-(The relations between these sub-generalizations embody phonological rules of generalizations embody phonological rules of the sort we have seen earlier)the sort we have seen earlier)

This phoneme is a much messier category than This phoneme is a much messier category than tt,, č č, and the others., and the others.

(… •) β (V…)[-bk]

(… •) ɸ (V…)[-bk]

{ }h

(… •) (V…)[+bk]

{ }h(… •) w (V…)

[+bk]

(… •) labial and/or velar

approximant/fricative (V…)(… •) w (V…)

[+bk][±vd]

(… •) β (V…)[-bk][±vd]

(… •) w/β (V…)

(… •) /ɸ (V…){ }h


In fact, In fact, ww is already such a messy category that is already such a messy category that adding adding hh to it may not seem like that big a to it may not seem like that big a change.change.

(This extension is the one we described as a sub-(This extension is the one we described as a sub-case of the Spirantization rule)case of the Spirantization rule)

(… •) β (V…)[-bk]

(… •) ɸ (V…)[-bk]

{ }h

(… •) (V…)[+bk]

{ }h(… •) w (V…)

[+bk]



[+bk][±vd]

(… •) β (V…)[-bk][±vd]

(… •) w/β (V…)

(… •) /ɸ (V…){ }h

(…) h ( •C…)


Still, it is less than clear what schema (if any) one Still, it is less than clear what schema (if any) one should posit as extracted from the extension.should posit as extracted from the extension.

Certainly that schema would not fit well into the Certainly that schema would not fit well into the nice set of phonemes in Mösiehualnice set of phonemes in Mösiehualii..

(… •) β (V…)[-bk]

(… •) ɸ (V…)[-bk]

{ }h

(… •) (V…)[+bk]

{ }h(… •) w (V…)

[+bk]



[+bk][±vd]

(… •) β (V…)[-bk][±vd]

(… •) w/β (V…)

(… •) /ɸ (V…){ }h

(…) h ( •C…)

(…) approximant /

fricative (…)

Neutralization to Neutralization to hh

More importantly (perhaps), this More importantly (perhaps), this hh “allophone” (or “allophone” (or particular subcases of it) is shared by other particular subcases of it) is shared by other phonemes. phonemes.

These phoneme categories do overlap.These phoneme categories do overlap.



Clearly distinguished subcases

w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y

Lessclearly distinguished

subcases

(…) h ( • …)(apico-dental stop)

t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š



What is more, these What is more, these h h “allophones” fit “allophones” fit quite well among the non-distinguished quite well among the non-distinguished allophones of another phoneme, namely allophones of another phoneme, namely hh..




w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y


subcases


t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š


h(unarticulated

pause in voicing)



What had looked like nice neat classical What had looked like nice neat classical categories suddenly aren’t—they categories suddenly aren’t—they overlap massively.overlap massively.




w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y


subcases


t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š


h(unarticulated

pause in voicing)



And the overlap is even more massive And the overlap is even more massive and complex, when we include the and complex, when we include the cases of cases of hh before before tt, , kk, , kʷkʷ, etc. , etc.




w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y


subcases


t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š


h(unarticulated

pause in voicing)


Consonant neutralizations Consonant neutralizations in Mösiehualin Mösiehualii

Not to mention the cases of neutralization Not to mention the cases of neutralization to to øø in word-final position. in word-final position.




w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y


subcases


t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š


h(unarticulated

pause in voicing)



All this may seem like making heavy weather of a All this may seem like making heavy weather of a sort of thing we’ve all seen plenty of times beforesort of thing we’ve all seen plenty of times before even if not quite on such a scale in most languages even if not quite on such a scale in most languages




w

(…) h ( •C…)

(palatal approx-

imant/fricative)

y


subcases


t


(alveolar

affricate)


(palato-alveolar

affricate)

č


(alveolar

fricative)

s


(…) h ( •č…)

(palato-alveolar

fricative)

š


h(unarticulated

pause in voicing)


Consonant Neutralizations Consonant Neutralizations in Mösiehualin Mösiehualii

Under at least some other theoretical Under at least some other theoretical models this sort of thing is essentially models this sort of thing is essentially inexplicable.inexplicable.

Many of us were taught to see this Many of us were taught to see this not as normal, but as pathological.not as normal, but as pathological.

If classical categories are the basis of If classical categories are the basis of phonology, why would you ever get phonology, why would you ever get this kind of a mess?this kind of a mess?

But, it is in fact quite normal, But, it is in fact quite normal, empirically, to get this kind of messempirically, to get this kind of mess

(even if, again, not usually on this (even if, again, not usually on this scale).scale).


The main points I would draw from all this are:The main points I would draw from all this are: Phonological categories are like other Phonological categories are like other

categories we find in language.categories we find in language. Like other categories, they are usage-Like other categories, they are usage-

based. based. It is usage (It is usage (along withalong with , not , not instead ofinstead of

brute “etic” similarity) that lets us know brute “etic” similarity) that lets us know which sounds (or other concepts) are to which sounds (or other concepts) are to be compared and categorized together.be compared and categorized together.

Like other categories, phonological Like other categories, phonological categories may be nice, neat and orderly categories may be nice, neat and orderly (“classical”), or they may be messy. (“classical”), or they may be messy.


Most of the time, they are at least a Most of the time, they are at least a little bit messy.little bit messy.

From the CG perspective, From the CG perspective, aberrations from the classical aberrations from the classical phonemic model are perfectly phonemic model are perfectly normal and to be expected.normal and to be expected.

What needs explaining is not why What needs explaining is not why the phonemic model isn’t perfect, the phonemic model isn’t perfect, but rather why it works so well in so but rather why it works so well in so many cases.many cases.


The CG perspective on categories The CG perspective on categories allows us to characterize, in natural and allows us to characterize, in natural and naturally related ways, such traditional naturally related ways, such traditional structures and mechanisms as:structures and mechanisms as:

Morphemic alternation rulesMorphemic alternation rules Morphophonemic and other phonological Morphophonemic and other phonological

rulesrules Rule orderingRule ordering Phonemes and allophonesPhonemes and allophones ArchiphonemesArchiphonemes FeaturesFeatures NeutralizationNeutralization


I think this is important. I think this is important. If a model obliges me to say that older If a model obliges me to say that older

thinkers were all wrong, something is thinkers were all wrong, something is probably wrong with the model. probably wrong with the model.

If it helps me understand what those If it helps me understand what those older thinkers were talking about, that older thinkers were talking about, that it is a point in its favor.it is a point in its favor.


And, finally, And, finally, Isn’t MösiehualIsn’t Mösiehualii beautiful? beautiful?

Isn’t it admirable how a community of Isn’t it admirable how a community of language-speakers can allow, and even language-speakers can allow, and even encourage to the point of enforcing, encourage to the point of enforcing, such massive neutralizations, such massive neutralizations,

and still be able to tell perfectly well and still be able to tell perfectly well what they are talking about?what they are talking about?

Power Point to be available atPower Point to be available at

www.sil.org/~tuggydwww.sil.org/~tuggyd

Neutralizations to h in Mösiehuali from the perspective of Cognitive Grammar David Tuggy SIL-Mexico.

Documents

c c otimokehc

h skirt

h stick

h life

squash h moayo

h baby mokon

skirt y mo

skirt n moc o